id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2407.13644 | Denys Nikolayev | Pratik Vadher, Anja K. Skrivervik, Qihang Zeng, Ronan Sauleau, John S. Ho, Giulia Sacco, Denys Nikolayev | Conformal Wide-Angle Scanning Leaky-Wave Antenna for V-Band On-Body Applications | null | null | null | null | physics.app-ph cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | Wearable on-body millimeter-wave (mmWave) radars can provide obstacle detection and guidance for visually impaired individuals. The antennas, being a crucial component of these systems, must be lightweight, flexible, low-cost, and compact. However, existing antennas suffer from a rigid form factor and limited reconfigurability. This article presents a low-profile, fast scanning leaky-wave antenna (LWA) operating in the unlicensed V-band (57-64 GHz) for on-body applications such as lightweight portable frequency modulated continuous wave (FMCW) radars. The novel meandering microstrip design allows independent control of gain and scanning rate (rate of change of main beam pointing direction with frequency). Experimental results show that the LWA achieves a realized gain above 10 dB with a fan-beam steering range in the H-plane from -35{deg} to 45{deg} over the operating frequency band, while the half power beamwidth (HPBW) is within 20{deg} in planar condition. To assess the on-body applicability, the antenna's performance is evaluated under bending. When placed on the knee (corresponding to 80 mm radius), the beam steers from -25{deg} to 55{deg} with a maximum realized gain degradation of 1.75 dB, and an increase of HPBW up to 25{deg}. This demonstrates the LWA's robustness in conformal conditions, while maintaining beam-forming and beam-scanning capabilities. Simulations confirm that the LWA's ground plane minimizes user exposure, adhering to international guidelines. Finally, we demonstrate a 2-D spatial scanning by employing an array of twelve LWAs with phased excitation, enabling beam-forming in the E-plane from -50{deg} to 50{deg}, while the HPBW remains below 20{deg}. Mutual coupling analysis reveals that isolation loss and active reflection coefficient remain below 15 dB throughout the operating band. | [
{
"version": "v1",
"created": "Thu, 18 Jul 2024 16:21:33 GMT"
},
{
"version": "v2",
"created": "Thu, 12 Sep 2024 09:18:42 GMT"
},
{
"version": "v3",
"created": "Tue, 27 May 2025 07:41:02 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Vadher",
"Pratik",
""
],
[
"Skrivervik",
"Anja K.",
""
],
[
"Zeng",
"Qihang",
""
],
[
"Sauleau",
"Ronan",
""
],
[
"Ho",
"John S.",
""
],
[
"Sacco",
"Giulia",
""
],
[
"Nikolayev",
"Denys",
""
]
] |
2407.14441 | Qinglin Tang | Weizhu Bao, Qinglin Tang, Yongjun Yuan | A normalized gradient flow method for computing ground states of spin-2 Bose-Einstein condensates | null | null | null | null | math.NA cond-mat.quant-gas cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose and analyze an efficient and accurate numerical method for computing ground states of spin-2 Bose-Einstein condensates (BECs) by using the normalized gradient flow (NGF). In order to successfully extend the NGF to spin-2 BECs which has five components in the vector wave function but with only two physical constraints on total mass conservation and magnetization conservation, two important techniques are introduced for designing the proposed numerical method. The first one is to systematically investigate the ground state structure and property of spin-2 BECs within a spatially uniform system, which can be used on how to properly choose initial data in the NGF for computing ground states of spin-2 BECs. The second one is to introduce three additional projection conditions based on the relations between the chemical potentials, together with the two existing physical constraints, such that the five projection parameters used in the projection step of the NGF can be uniquely determined. Then a backward-forward Euler finite difference method is adapted to discretize the NGF. We prove rigorously that there exists a unique solution of the nonlinear system for determining the five projection parameters in the full discretization of the NGF under a mild condition on the time step size. Extensive numerical results on ground states of spin-2 BECs with different types of phases and under different potentials are reported to show the efficiency and accuracy of the proposed numerical method and to demonstrate several interesting physical phenomena on ground states of spin-2 BECs. | [
{
"version": "v1",
"created": "Fri, 19 Jul 2024 16:13:59 GMT"
},
{
"version": "v2",
"created": "Wed, 26 Mar 2025 01:15:52 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 10:15:37 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Bao",
"Weizhu",
""
],
[
"Tang",
"Qinglin",
""
],
[
"Yuan",
"Yongjun",
""
]
] |
2407.14624 | Spencer Hamblen | Lucas Anthony, Joe Burton, Irene Deegbe, Sarah England, Spencer Hamblen, Reagan Knowles, Sarah Stewart, Hannah Wright | Waring Numbers of Ramified $p$-adic Rings | 21 pages, 5 figures | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $p$ be prime and $e,k$ be positive integers, and let $R = {\mathbb Z}_p[\sqrt[e]{p}]$. We calculate the Waring numbers $g_R(k)$ for many values of $p, e$, and $k$, and investigate how the Waring numbers for $p=2$ change as $e$ and $k$ vary. | [
{
"version": "v1",
"created": "Fri, 19 Jul 2024 18:49:02 GMT"
},
{
"version": "v2",
"created": "Sun, 01 Sep 2024 01:05:55 GMT"
},
{
"version": "v3",
"created": "Tue, 18 Mar 2025 18:42:53 GMT"
},
{
"version": "v4",
"created": "Sat, 07 Jun 2025 22:21:18 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Anthony",
"Lucas",
""
],
[
"Burton",
"Joe",
""
],
[
"Deegbe",
"Irene",
""
],
[
"England",
"Sarah",
""
],
[
"Hamblen",
"Spencer",
""
],
[
"Knowles",
"Reagan",
""
],
[
"Stewart",
"Sarah",
""
],
[
"Wright",
"Hannah",
""
]
] |
2407.15239 | Mariya Hendriksen | Mariya Hendriksen, Shuo Zhang, Ridho Reinanda, Mohamed Yahya, Edgar Meij, and Maarten de Rijke | Benchmark Granularity and Model Robustness for Image-Text Retrieval | accepted at SIGIR 2025 | null | null | null | cs.CV cs.AI cs.IR | http://creativecommons.org/licenses/by/4.0/ | Image-Text Retrieval (ITR) systems are central to multimodal information access, with Vision-Language Models (VLMs) showing strong performance on standard benchmarks. However, these benchmarks predominantly rely on coarse-grained annotations, limiting their ability to reveal how models perform under real-world conditions, where query granularity varies. Motivated by this gap, we examine how dataset granularity and query perturbations affect retrieval performance and robustness across four architecturally diverse VLMs (ALIGN, AltCLIP, CLIP, and GroupViT). Using both standard benchmarks (MS-COCO, Flickr30k) and their fine-grained variants, we show that richer captions consistently enhance retrieval, especially in text-to-image tasks, where we observe an average improvement of 16.23%, compared to 6.44% in image-to-text. To assess robustness, we introduce a taxonomy of perturbations and conduct extensive experiments, revealing that while perturbations typically degrade performance, they can also unexpectedly improve retrieval, exposing nuanced model behaviors. Notably, word order emerges as a critical factor -- contradicting prior assumptions of model insensitivity to it. Our results highlight variation in model robustness and a dataset-dependent relationship between caption granularity and perturbation sensitivity and emphasize the necessity of evaluating models on datasets of varying granularity. | [
{
"version": "v1",
"created": "Sun, 21 Jul 2024 18:08:44 GMT"
},
{
"version": "v2",
"created": "Thu, 25 Jul 2024 19:52:38 GMT"
},
{
"version": "v3",
"created": "Mon, 28 Oct 2024 17:52:17 GMT"
},
{
"version": "v4",
"created": "Mon, 09 Jun 2025 14:08:47 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Hendriksen",
"Mariya",
""
],
[
"Zhang",
"Shuo",
""
],
[
"Reinanda",
"Ridho",
""
],
[
"Yahya",
"Mohamed",
""
],
[
"Meij",
"Edgar",
""
],
[
"de Rijke",
"Maarten",
""
]
] |
2407.15547 | Jasson Vindas | Morgan Callewaert, Lenny Neyt, Jasson Vindas | On a space of functions with entire Laplace transforms and its connection with the optimality of the Ingham-Karamata theorem | 11 pages, 2 figures | Complex Anal. Oper. Theory 19 (2025), Article 85 | 10.1007/s11785-025-01704-8 | null | math.CA math.CV math.FA math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study approximation properties of the Fr\'{e}chet space of all continuously differentiable functions $\tau$ such that $\tau'(x)=o(1)$ and such that their Laplace transforms admit entire extensions to $\mathbb{C}$. As an application, these approximation results are combined with the open mapping theorem to show the optimality theorem for the Ingham-Karamata Tauberian theorem. | [
{
"version": "v1",
"created": "Mon, 22 Jul 2024 11:14:31 GMT"
},
{
"version": "v2",
"created": "Tue, 20 May 2025 15:31:08 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Callewaert",
"Morgan",
""
],
[
"Neyt",
"Lenny",
""
],
[
"Vindas",
"Jasson",
""
]
] |
2407.15710 | Henning Korbmacher | Henning Korbmacher, Gustavo A. Dom\'inguez-Castro, Mateusz {\L}\k{a}cki, Jakub Zakrzewski, Luis Santos | Topological floating phase of dipolar bosons in an optical ladder | null | null | null | null | cond-mat.quant-gas | http://creativecommons.org/licenses/by/4.0/ | Ultracold dipolar hard-core bosons in optical ladders provide exciting possibilities for the quantum simulation of anisotropic XXZ spin ladders. We show that introducing a tilt along the rungs results in a rich phase diagram at unit filling. In particular, for a sufficiently strong dipolar strength, the interplay between the long-range tail of the dipolar interactions and the tilting leads to the emergence of a quantum floating phase, a critical phase with incommensurate density-density correlations. Interestingly, the study of the entanglement spectrum, reveals that the floating phase is topological, constituting an intermediate gapless stage in the melting of a crystal into a gapped topological Haldane phase. This novel scenario for topological floating phases in dipolar XXZ ladders can be investigated in on-going experiments. | [
{
"version": "v1",
"created": "Mon, 22 Jul 2024 15:18:15 GMT"
},
{
"version": "v2",
"created": "Thu, 05 Sep 2024 12:51:37 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Feb 2025 16:13:00 GMT"
},
{
"version": "v4",
"created": "Sat, 07 Jun 2025 08:39:44 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Korbmacher",
"Henning",
""
],
[
"Domínguez-Castro",
"Gustavo A.",
""
],
[
"Łącki",
"Mateusz",
""
],
[
"Zakrzewski",
"Jakub",
""
],
[
"Santos",
"Luis",
""
]
] |
2407.16363 | Josephine Hunout | Josephine Hunout, Sylvain Laizet and Lorenzo Iannucci | Variational quantum algorithm based on Lagrange polynomial encoding to solve differential equations | 21 pages, 19 figures | Phys. Rev. A, vol. 111, no. 6, p. 062404, Jun. 2025 | 10.1103/PhysRevA.111.062404 | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Differential equations (DEs) serve as the cornerstone for a wide range of scientific endeavors, their solutions weaving through the core of diverse fields such as structural engineering, fluid dynamics, and financial modeling. DEs are notoriously hard to solve, due to their intricate nature, and finding solutions to DEs often exceeds the capabilities of traditional computational approaches. Recent advances in quantum computing have triggered a growing interest from researchers for the design of quantum algorithms for solving DEs. In this work, we introduce two different architectures of a novel variational quantum algorithm (VQA) with Lagrange polynomial encoding in combination with derivative quantum circuits using the Hadamard test differentiation to approximate the solution of DEs. To demonstrate the potential of our new VQA, two well-known ordinary differential equations are used: the damped mass-spring system from a given initial condition and the Poisson equation for periodic, Dirichlet, and Neumann boundary conditions. It is shown that the proposed new VQA has a reduced gate complexity compared to previous variational quantum algorithms, for a similar or better quality of the solution. | [
{
"version": "v1",
"created": "Tue, 23 Jul 2024 10:11:44 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Apr 2025 18:31:13 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 17:01:09 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Hunout",
"Josephine",
""
],
[
"Laizet",
"Sylvain",
""
],
[
"Iannucci",
"Lorenzo",
""
]
] |
2407.16372 | L\'eon Burkhardt | L\'eon Burkhardt | Weight 11 Compactly Supported Cohomology of Moduli Spaces of Curves in excess four | 13 pages [edit] minor typos corrected, to appear in Homology, Homotopy and Applications | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In their homonymous article, Sam Payne and Thomas Willwacher construct a combinatorial graph complex to compute the weight 11 part of the compactly supported cohomology of the moduli space of curves $\cM_{g,n}$ and compute explicitly the cohomology of the introduced graph complexes in cases of complexes of excess zero, one, two, and three. In this paper, we extend the computation the cohomology to excess four graph complexes. Along the way, we give more details on generators, the computation process, and the definition of the graph complex. | [
{
"version": "v1",
"created": "Tue, 23 Jul 2024 10:42:01 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Dec 2024 17:16:29 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 07:58:09 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Burkhardt",
"Léon",
""
]
] |
2407.16803 | Abhi Kamboj | Abhi Kamboj, Anh Duy Nguyen, Minh N. Do | C3T: Cross-modal Transfer Through Time for Sensor-based Human Activity Recognition | null | null | null | null | cs.CV cs.AI cs.HC cs.LG eess.SP | http://creativecommons.org/licenses/by-sa/4.0/ | In order to unlock the potential of diverse sensors, we investigate a method to transfer knowledge between time-series modalities using a multimodal \textit{temporal} representation space for Human Activity Recognition (HAR). Specifically, we explore the setting where the modality used in testing has no labeled data during training, which we refer to as Unsupervised Modality Adaptation (UMA). We categorize existing UMA approaches as Student-Teacher or Contrastive Alignment methods. These methods typically compress continuous-time data samples into single latent vectors during alignment, inhibiting their ability to transfer temporal information through real-world temporal distortions. To address this, we introduce Cross-modal Transfer Through Time (C3T), which preserves temporal information during alignment to handle dynamic sensor data better. C3T achieves this by aligning a set of temporal latent vectors across sensing modalities. Our extensive experiments on various camera+IMU datasets demonstrate that C3T outperforms existing methods in UMA by at least 8% in accuracy and shows superior robustness to temporal distortions such as time-shift, misalignment, and dilation. Our findings suggest that C3T has significant potential for developing generalizable models for time-series sensor data, opening new avenues for various multimodal applications. | [
{
"version": "v1",
"created": "Tue, 23 Jul 2024 19:06:44 GMT"
},
{
"version": "v2",
"created": "Thu, 07 Nov 2024 17:10:15 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 15:03:39 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kamboj",
"Abhi",
""
],
[
"Nguyen",
"Anh Duy",
""
],
[
"Do",
"Minh N.",
""
]
] |
2407.17493 | Youngseok Yoon | Youngseok Yoon, Dainong Hu, Iain Weissburg, Yao Qin, Haewon Jeong | Model Collapse in the Self-Consuming Chain of Diffusion Finetuning: A Novel Perspective from Quantitative Trait Modeling | version 3, additional 34 pages | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Model collapse, the severe degradation of generative models when iteratively trained on their own outputs, has gained significant attention in recent years. This paper examines Chain of Diffusion, where a pretrained text-to-image diffusion model is finetuned on its own generated images. We demonstrate that severe image quality degradation was universal and identify CFG scale as the key factor impacting this model collapse. Drawing on an analogy between the Chain of Diffusion and biological evolution, we then introduce a novel theoretical analysis based on quantitative trait modeling from statistical genetics. Our theoretical analysis aligns with empirical observations of the generated images in the Chain of Diffusion. Finally, we propose Reusable Diffusion Finetuning (ReDiFine), a simple yet effective strategy inspired by genetic mutations. It operates robustly across various scenarios without requiring any hyperparameter tuning, making it a plug-and-play solution for reusable image generation. | [
{
"version": "v1",
"created": "Thu, 04 Jul 2024 13:41:54 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Oct 2024 20:03:46 GMT"
},
{
"version": "v3",
"created": "Fri, 06 Jun 2025 22:43:49 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yoon",
"Youngseok",
""
],
[
"Hu",
"Dainong",
""
],
[
"Weissburg",
"Iain",
""
],
[
"Qin",
"Yao",
""
],
[
"Jeong",
"Haewon",
""
]
] |
2407.17929 | Krishnakant Singh | Krishnakant Singh, Simone Schaub-Meyer, Stefan Roth | GLASS: Guided Latent Slot Diffusion for Object-Centric Learning | CVPR 2025. Project Page: http://visinf.github.io/glass/ | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Object-centric learning aims to decompose an input image into a set of meaningful object files (slots). These latent object representations enable a variety of downstream tasks. Yet, object-centric learning struggles on real-world datasets, which contain multiple objects of complex textures and shapes in natural everyday scenes. To address this, we introduce Guided Latent Slot Diffusion (GLASS), a novel slot-attention model that learns in the space of generated images and uses semantic and instance guidance modules to learn better slot embeddings for various downstream tasks. Our experiments show that GLASS surpasses state-of-the-art slot-attention methods by a wide margin on tasks such as (zero-shot) object discovery and conditional image generation for real-world scenes. Moreover, GLASS enables the first application of slot attention to the compositional generation of complex, realistic scenes. | [
{
"version": "v1",
"created": "Thu, 25 Jul 2024 10:38:32 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 03:45:15 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Singh",
"Krishnakant",
""
],
[
"Schaub-Meyer",
"Simone",
""
],
[
"Roth",
"Stefan",
""
]
] |
2407.18152 | Jiacheng Sun | Jia-Cheng Sun, Chi Zhang, Haoran Zhu | Kronecker coefficients and Harrison centres of the representation ring of the symmetric group | 21 pages; revised version | null | null | null | math.RT math.RA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a computational approach to studying the structure of the representation ring of the symmetric group in dimension six. The Kronecker coefficients and all power formulae of irreducible representations of $S_6$ are computed using the character theory of finite groups. In addition, considering direct sum decomposition of tensor products of different irreducible representations of $S_6$, we characterise generators of the representation ring $\mathcal{R}(S_6)$, show that its unit group $U(\mathcal{R}(S_6))$ is a Klein four-group, and related results on the structure of primitive idempotents. Furthermore, we introduce the Harrison centre theory to study the representation ring and show that the Harrison centre of the cubic form induced by the generating relations of $\mathcal{R}(S_6)$ is isomorphic to itself. Finally, we conclude with some open problems for future consideration. | [
{
"version": "v1",
"created": "Thu, 25 Jul 2024 16:03:12 GMT"
},
{
"version": "v2",
"created": "Tue, 06 Aug 2024 01:10:29 GMT"
},
{
"version": "v3",
"created": "Mon, 06 Jan 2025 12:54:02 GMT"
},
{
"version": "v4",
"created": "Sun, 08 Jun 2025 12:52:47 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Sun",
"Jia-Cheng",
""
],
[
"Zhang",
"Chi",
""
],
[
"Zhu",
"Haoran",
""
]
] |
2407.18296 | Nudzeim Selimovic | Jorge Alda, Gabriele Levati, Paride Paradisi, Stefano Rigolin, Nudzeim Selimovic | Collider and astrophysical signatures of light scalars with enhanced $\tau$ couplings | 9 pages, 6 figures, v2: matches published version | null | 10.1007/JHEP06(2025)008 | null | hep-ph hep-ex | http://creativecommons.org/licenses/by/4.0/ | Beyond Standard Model scenarios addressing the flavor puzzle and the hierarchy problem generally predict dominant new physics couplings with fermions of the third generation. In this Letter, we explore the collider and astrophysical signatures of new light scalar and pseudoscalar particles dominantly coupled to the $\tau$-lepton. The best experimental prospects are expected at Belle II through the $e^+e^-\to\tau^+\tau^-\gamma\gamma$, $\tau^+\tau^-\gamma$, $3\gamma$, mono-$\gamma$ processes, and the $\tau$ anomalous magnetic moment. The correlated effects in these searches can unambiguously point toward the underlying new physics dynamics. Moreover, we study astrophysics bounds - especially from core-collapse supernovae and neutron star mergers - finding them particularly effective and complementary to collider bounds. We carry out this program in the well-motivated context of axion-like particles as well as generic CP-even and CP-odd particles, highlighting possible ways to discriminate among them. | [
{
"version": "v1",
"created": "Thu, 25 Jul 2024 18:00:01 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 10:21:51 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Alda",
"Jorge",
""
],
[
"Levati",
"Gabriele",
""
],
[
"Paradisi",
"Paride",
""
],
[
"Rigolin",
"Stefano",
""
],
[
"Selimovic",
"Nudzeim",
""
]
] |
2407.18950 | Jongwoo Seo | Jongwoo Seo | Unexplainability of Artificial Intelligence Judgments in Kant's Perspective | 8 pages, 1 figure | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Kant's Critique of Pure Reason, a major contribution to the history of epistemology, proposes a table of categories to elucidate the structure of the a priori principles underlying human judgment. Artificial intelligence (AI) technology, grounded in functionalism, claims to simulate or replicate human judgment. To evaluate this claim, it is necessary to examine whether AI judgments exhibit the essential characteristics of human judgment. This paper investigates the unexplainability of AI judgments through the lens of Kant's theory of judgment. Drawing on Kant's four logical forms-quantity, quality, relation, and modality-this study identifies what may be called AI's uncertainty, a condition in which different forms of judgment become entangled. In particular, with regard to modality, this study argues that the SoftMax function forcibly reframes AI judgments as possibility judgments. Furthermore, since complete definitions in natural language are impossible, words are, by their very nature, ultimately unexplainable; therefore, a fully complete functional implementation is theoretically unattainable. | [
{
"version": "v1",
"created": "Fri, 12 Jul 2024 03:39:55 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Sep 2024 14:33:08 GMT"
},
{
"version": "v3",
"created": "Sat, 31 May 2025 01:31:03 GMT"
},
{
"version": "v4",
"created": "Mon, 09 Jun 2025 13:12:25 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Seo",
"Jongwoo",
""
]
] |
2407.19326 | Birte Boes | Birte Boes, Jaan-Willem Simon, Hagen Holthusen | Accounting for plasticity: An extension of inelastic Constitutive Artificial Neural Networks | 36 pages, 10 figures, 5 tables | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | In this work, we extend the existing framework of inelastic constitutive artificial neural networks (iCANNs) by incorporating plasticity to increase their applicability to model more complex material behavior. The proposed approach ensures objectivity, material symmetry, and thermodynamic consistency, providing a robust basis for automatic model discovery of constitutive equations at finite strains. These are predicted by discovering formulations for the Helmholtz free energy and plastic potentials for the yield function and evolution equations in terms of feed-forward networks. Our framework captures both linear and nonlinear kinematic hardening behavior. Investigation of our model's prediction showed that the extended iCANNs successfully predict both linear and nonlinear kinematic hardening behavior based on experimental and artificially generated datasets, showcasing promising capabilities of this framework. Nonetheless, challenges remain in discovering more complex yield criteria with tension-compression asymmetry and addressing deviations in experimental data at larger strains. Despite these limitations, the proposed framework provides a promising basis for incorporating plasticity into iCANNs, offering a platform for advancing in the field of automated model discovery. | [
{
"version": "v1",
"created": "Sat, 27 Jul 2024 19:19:42 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 06:08:04 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Boes",
"Birte",
""
],
[
"Simon",
"Jaan-Willem",
""
],
[
"Holthusen",
"Hagen",
""
]
] |
2407.19353 | Cheng Shi | Cheng Shi, Liming Pan and Ivan Dokmani\'c | A spring-block theory of feature learning in deep neural networks | null | null | null | null | cond-mat.dis-nn cond-mat.stat-mech cs.LG stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Feature-learning deep nets progressively collapse data to a regular low-dimensional geometry. How this emerges from the collective action of nonlinearity, noise, learning rate, and other factors, has eluded first-principles theories built from microscopic neuronal dynamics. We exhibit a noise-nonlinearity phase diagram that identifies regimes where shallow or deep layers learn more effectively and propose a macroscopic mechanical theory that reproduces the diagram and links feature learning across layers to generalization. | [
{
"version": "v1",
"created": "Sun, 28 Jul 2024 00:07:20 GMT"
},
{
"version": "v2",
"created": "Wed, 23 Oct 2024 14:11:34 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 17:25:08 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Shi",
"Cheng",
""
],
[
"Pan",
"Liming",
""
],
[
"Dokmanić",
"Ivan",
""
]
] |
2407.19807 | Cong Liu | Cong Liu, Xiaojun Quan, Yan Pan, Liang Lin, Weigang Wu, Xu Chen | Cool-Fusion: Fuse Large Language Models without Training | null | ACL 2025 | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We focus on the problem of fusing two or more heterogeneous large language models (LLMs) to leverage their complementary strengths. One of the challenges of model fusion is high computational load, specifically in fine-tuning or aligning vocabularies. To address this, we propose Cool-Fusion, a simple yet effective approach that fuses the knowledge of source LLMs, which does not require training. Unlike ensemble methods, Cool-Fusion is applicable to any set of source LLMs that have different vocabularies. To overcome the vocabulary discrepancies among LLMs, we ensemble LLMs on text level, allowing them to rerank the generated texts by each other with different granularities. Extensive experiments have been conducted across a variety of benchmark datasets. On GSM8K, Cool-Fusion increases accuracy from three strong source LLMs by a significant margin of 17.4\%. | [
{
"version": "v1",
"created": "Mon, 29 Jul 2024 09:02:19 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 09:19:19 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Liu",
"Cong",
""
],
[
"Quan",
"Xiaojun",
""
],
[
"Pan",
"Yan",
""
],
[
"Lin",
"Liang",
""
],
[
"Wu",
"Weigang",
""
],
[
"Chen",
"Xu",
""
]
] |
2407.21340 | Aydogan Ozcan | Yuntian Wang, Yuhang Li, Tianyi Gan, Kun Liao, Mona Jarrahi, Aydogan Ozcan | Optimizing Structured Surfaces for Diffractive Waveguides | 60 Pages, 14 Figures | Nature Communications (2025) | 10.1038/s41467-025-60626-3 | null | physics.optics physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Waveguide design is crucial in developing efficient light delivery systems, requiring meticulous material selection, precise manufacturing, and rigorous performance optimization, including dispersion engineering. Here, we introduce universal diffractive waveguide designs that can match the performance of any conventional dielectric waveguide and achieve various functionalities. Optimized using deep learning, our diffractive waveguide designs can be cascaded to each other to form any desired length and are comprised of transmissive diffractive surfaces that permit the propagation of desired guided modes with low loss and high mode purity. In addition to guiding the targeted modes along the propagation direction through cascaded diffractive units, we also developed various waveguide components and introduced bent diffractive waveguides, rotating the direction of mode propagation, as well as spatial and spectral mode filtering and mode splitting diffractive waveguide designs, and mode-specific polarization-maintaining diffractive waveguides, showcasing the versatility of this platform. This diffractive waveguide framework was experimentally validated in the terahertz (THz) spectrum using 3D-printed diffractive layers to selectively pass certain spatial modes while rejecting others. Diffractive waveguides can be scaled to operate at different wavelengths, including visible and infrared parts of the spectrum, without the need for material dispersion engineering, providing an alternative to traditional waveguide components. This advancement will have potential applications in telecommunications, imaging, sensing and spectroscopy, among others. | [
{
"version": "v1",
"created": "Wed, 31 Jul 2024 05:14:44 GMT"
},
{
"version": "v2",
"created": "Mon, 07 Apr 2025 05:11:51 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wang",
"Yuntian",
""
],
[
"Li",
"Yuhang",
""
],
[
"Gan",
"Tianyi",
""
],
[
"Liao",
"Kun",
""
],
[
"Jarrahi",
"Mona",
""
],
[
"Ozcan",
"Aydogan",
""
]
] |
2407.21698 | Ning Qi | Ning Qi, Kaidi Huang, Zhiyuan Fan and Bolun Xu | Long-Term Energy Management for Microgrid with Hybrid Hydrogen-Battery Energy Storage: A Prediction-Free Coordinated Optimization Framework | Submitted to Applied Energy | Applied Energy, vol. 377, p. 124485, 2025 | 10.1016/j.apenergy.2024.124485 | null | math.OC cs.SY eess.SY | http://creativecommons.org/licenses/by/4.0/ | This paper studies the long-term energy management of a microgrid coordinating hybrid hydrogen-battery energy storage. We develop an approximate semi-empirical hydrogen storage model to accurately capture the power-dependent efficiency of hydrogen storage. We introduce a prediction-free two-stage coordinated optimization framework, which generates the annual state-of-charge (SoC) reference for hydrogen storage offline. During online operation, it updates the SoC reference online using kernel regression and makes operation decisions based on the proposed adaptive virtual-queue-based online convex optimization (OCO) algorithm. We innovatively incorporate penalty terms for long-term pattern tracking and expert-tracking for step size updates. We provide theoretical proof to show that the proposed OCO algorithm achieves a sublinear bound of dynamic regret without using prediction information. Numerical studies based on the Elia and North China datasets show that the proposed framework significantly outperforms the existing online optimization approaches by reducing the operational costs and loss of load by around 30% and 80%, respectively. These benefits can be further enhanced with optimized settings for the penalty coefficient and step size of OCO, as well as more historical references. | [
{
"version": "v1",
"created": "Wed, 31 Jul 2024 15:44:07 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Qi",
"Ning",
""
],
[
"Huang",
"Kaidi",
""
],
[
"Fan",
"Zhiyuan",
""
],
[
"Xu",
"Bolun",
""
]
] |
2408.00064 | Till Sawala | Till Sawala (1 and 2), Jehanne Delhomelle (1 and 3), Alis J. Deason (2), Carlos S. Frenk (2), Peter H. Johansson (1), Atte Keitaanranta (1), Alexander Rawlings (1), Ruby Wright (1) ((1) University of Helsinki, (2) Durham University, (3) University of Toulouse III - Paul Sabatier) | Apocalypse When? No Certainty of a Milky Way -- Andromeda Collision | null | null | 10.1038/s41550-025-02563-1 | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | It is commonly believed that our own Milky Way is on a collision course with the neighbouring Andromeda galaxy. As a result of their merger, predicted in around five billion years, the two large spiral galaxies that define the present Local Group would form a new elliptical galaxy. Here we consider the latest and most accurate observations by the Gaia and Hubble space telescopes, along with recent consensus mass estimates to derive possible future scenarios and identify the major sources of uncertainty in the evolution of the Local Group over the next 10 billion years. We find that the next most massive Local Group member galaxies -- namely, M33 and the Large Magellanic Cloud -- distinctly and radically affect the Milky Way - Andromeda orbit. While including M33 increases the merger probability, the orbit of the Large Magellanic Cloud runs perpendicular to the Milky Way - Andromeda orbit and makes their merger less likely. In the full system, we find that uncertainties in the present positions, motions, and masses of all galaxies leave room for drastically different outcomes, and a probability of close to 50% that there is no Milky Way - Andromeda merger during the next 10 billion years. | [
{
"version": "v1",
"created": "Wed, 31 Jul 2024 18:00:00 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Sawala",
"Till",
"",
"1 and 2"
],
[
"Delhomelle",
"Jehanne",
"",
"1 and 3"
],
[
"Deason",
"Alis J.",
"",
"Durham University"
],
[
"Frenk",
"Carlos S.",
"",
"Durham University"
],
[
"Johansson",
"Peter H.",
"",
"University of Helsinki"
],
[
"Keitaanranta",
"Atte",
"",
"University of Helsinki"
],
[
"Rawlings",
"Alexander",
"",
"University of Helsinki"
],
[
"Wright",
"Ruby",
"",
"University of Helsinki"
]
] |
2408.00926 | Kazuhiko Seki | Kazuhiko Seki | Method of images for one-dimensional discrete random walk under a reflecting barrier | 7 pages: 1 figure | null | null | null | cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | The transition probability for a one-dimensional discrete symmetric random walk under a reflecting barrier was once given by the method of images. [S. Chandrasekhar, Rev. Mod. Phys. 15, 1 (1943).] However, several inconsistencies have been reported when the method of images is applied in cases where a reflecting barrier is considered, even after the exact solution has been obtained. Here, we explicitly show that the method of images becomes applicable if the image position is shifted. | [
{
"version": "v1",
"created": "Thu, 01 Aug 2024 21:28:40 GMT"
},
{
"version": "v2",
"created": "Sun, 18 May 2025 06:50:36 GMT"
},
{
"version": "v3",
"created": "Fri, 06 Jun 2025 05:49:30 GMT"
},
{
"version": "v4",
"created": "Mon, 09 Jun 2025 12:08:41 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Seki",
"Kazuhiko",
""
]
] |
2408.01076 | Lu Yu | Lu Yu, Zhe Tao, Dipam Goswami, Hantao Yao, Bart{\l}omiej Twardowski, Joost Van de Weijer and Changsheng Xu | Exploiting the Semantic Knowledge of Pre-trained Text-Encoders for Continual Learning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Deep neural networks (DNNs) excel on fixed datasets but struggle with incremental and shifting data in real-world scenarios. Continual learning addresses this challenge by allowing models to learn from new data while retaining previously learned knowledge. Existing methods mainly rely on visual features, often neglecting the rich semantic information encoded in text. The semantic knowledge available in the label information of the images, offers important semantic information that can be related with previously acquired knowledge of semantic classes. Consequently, effectively leveraging this information throughout continual learning is expected to be beneficial. To address this, we propose integrating semantic guidance within and across tasks by capturing semantic similarity using text embeddings. We start from a pre-trained CLIP model, employ the \emph{Semantically-guided Representation Learning (SG-RL)} module for a soft-assignment towards all current task classes, and use the Semantically-guided Knowledge Distillation (SG-KD) module for enhanced knowledge transfer. Experimental results demonstrate the superiority of our method on general and fine-grained datasets. Our code can be found in https://github.com/aprilsveryown/semantically-guided-continual-learning. | [
{
"version": "v1",
"created": "Fri, 02 Aug 2024 07:51:44 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 08:48:46 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yu",
"Lu",
""
],
[
"Tao",
"Zhe",
""
],
[
"Goswami",
"Dipam",
""
],
[
"Yao",
"Hantao",
""
],
[
"Twardowski",
"Bartłomiej",
""
],
[
"Van de Weijer",
"Joost",
""
],
[
"Xu",
"Changsheng",
""
]
] |
2408.01583 | Michael Wentzel | Daniel Carney, Gerard Higgins, Giacomo Marocco, Michael Wentzel | Superconducting Levitated Detector of Gravitational Waves | 5 pages, 2 pages of end matter, 10 pages of supplemental material | Carney (2025) Phys. Rev. Lett. 134, 181402 | 10.1103/PhysRevLett.134.181402 | null | hep-ph astro-ph.HE gr-qc hep-ex | http://creativecommons.org/licenses/by/4.0/ | A magnetically levitated mass couples to gravity and can act as an effective gravitational wave detector. We show that a superconducting sphere levitated in a quadrupolar magnetic field, when excited by a gravitational wave, will produce magnetic field fluctuations that can be read out using a flux tunable microwave resonator. With a readout operating at the standard quantum limit, such a system could achieve broadband strain noise sensitivity of $h \lesssim 10^{-20}/\sqrt{\rm Hz}$ for frequencies of $1~\mathrm{kHz}~-~1~\mathrm{MHz}$, opening new corridors for astrophysical probes of new physics. | [
{
"version": "v1",
"created": "Fri, 02 Aug 2024 21:36:26 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 19:18:11 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Carney",
"Daniel",
""
],
[
"Higgins",
"Gerard",
""
],
[
"Marocco",
"Giacomo",
""
],
[
"Wentzel",
"Michael",
""
]
] |
2408.02351 | Orlando Luongo | Alessio Lapponi, Orlando Luongo, Stefano Mancini | How cosmological expansion affects communication between distant quantum systems | 22 pages, 4 figures | Phys.Rev.D 111 (2025) 4, 044040 | 10.1103/PhysRevD.111.044040 | null | gr-qc astro-ph.CO quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A quantum communication protocol between harmonic oscillator detectors, interacting with a quantum field, is developed in a cosmological expanding background. The aim is to see if the quantum effects arising in an expanding universe, such as the cosmological particle production, could facilitate the communication between two distant parts or if they provide an additive noisy effect. By considering a perfect cosmic fluid, the resulting expansion turns out to increase the classical capacity of the protocol. This increasing occurs for all the cosmological expansions unless the latter is sharpened just before the receiver's detector interacts with the field. Moreover, the classical capacity turns out to be sensible to the barotropic parameter $w$ of the perfect fluid and to the coupling between the field and the scalar curvature $\xi$. As a consequence, by performing this protocol, one can achieve information about the cosmological dynamics and its coupling with a background quantum field. | [
{
"version": "v1",
"created": "Mon, 05 Aug 2024 09:59:27 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Lapponi",
"Alessio",
""
],
[
"Luongo",
"Orlando",
""
],
[
"Mancini",
"Stefano",
""
]
] |
2408.02902 | Yunyan Yang | Mengjie Zhang, Yong Lin, Yunyan Yang | Fractional Laplace operator and related Schr\"odinger equations on locally finite graphs | 24 pages | null | null | null | math.AP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we first define a discrete version of the fractional Laplace operator $(-\Delta)^{s}$ through the heat semigroup on a stochastically complete, connected, locally finite graph $G = (V, E, \mu, w)$. Secondly, we define the fractional divergence and give another form of $(-\Delta)^s$. The third point, and the foremost, is the introduction of the fractional Sobolev space $W^{s,2}(V)$, which is necessary when we study problems involving $(-\Delta)^{s}$. Finally, using the mountain-pass theorem and the Nehari manifold, we obtain multiplicity solutions to a discrete fractional Schr\"{o}dinger equation on $G$. We caution the readers that though these existence results are well known in the continuous case, the discrete case is quite different. | [
{
"version": "v1",
"created": "Tue, 06 Aug 2024 02:22:06 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 10:46:16 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhang",
"Mengjie",
""
],
[
"Lin",
"Yong",
""
],
[
"Yang",
"Yunyan",
""
]
] |
2408.03265 | Jakub Nawa{\l}a | Jakub Nawa{\l}a, Yuxuan Jiang, Fan Zhang, Xiaoqing Zhu, Joel Sole, David Bull | BVI-AOM: A New Training Dataset for Deep Video Compression Optimization | 5 pages, 5 figures. Swapped the PSNR-HVS plot in Fig. 3 for a PSNR-YUV plot. Updated Fig. 3 (SI/TI/CF plots) and added the URL to the dataset | null | 10.1109/VCIP63160.2024.10849837 | null | eess.IV | http://creativecommons.org/licenses/by/4.0/ | Deep learning is now playing an important role in enhancing the performance of conventional hybrid video codecs. These learning-based methods typically require diverse and representative training material for optimization in order to achieve model generalization and optimal coding performance. However, existing datasets either offer limited content variability or come with restricted licensing terms constraining their use to research purposes only. To address these issues, we propose a new training dataset, named BVI-AOM, which contains 956 uncompressed sequences at various resolutions from 270p to 2160p, covering a wide range of content and texture types. The dataset comes with more flexible licensing terms and offers competitive performance when used as a training set for optimizing deep video coding tools. The experimental results demonstrate that when used as a training set to optimize two popular network architectures for two different coding tools, the proposed dataset leads to additional bitrate savings of up to 0.29 and 2.98 percentage points in terms of PSNR-Y and VMAF, respectively, compared to an existing training dataset, BVI-DVC, which has been widely used for deep video coding. The BVI-AOM dataset is available at https://github.com/fan-aaron-zhang/bvi-aom | [
{
"version": "v1",
"created": "Tue, 06 Aug 2024 15:54:55 GMT"
},
{
"version": "v2",
"created": "Wed, 07 Aug 2024 17:39:17 GMT"
},
{
"version": "v3",
"created": "Wed, 23 Oct 2024 09:34:55 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Nawała",
"Jakub",
""
],
[
"Jiang",
"Yuxuan",
""
],
[
"Zhang",
"Fan",
""
],
[
"Zhu",
"Xiaoqing",
""
],
[
"Sole",
"Joel",
""
],
[
"Bull",
"David",
""
]
] |
2408.03364 | T\'essio B. De Melo | Carolina Arbel\'aez, Giovanna Cottin, Juan Carlos Helo, Martin Hirsch, T\'essio B. de Melo | Long-lived particle phenomenology in one-loop neutrino mass models with dark matter | 23 pages, 8 figures, 1 table. Matches version published in JHEP | null | 10.1007/JHEP02(2025)049 | null | hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neutrino masses and dark matter (DM) might have a common origin. The scotogenic model can be considered the proto-type model realizing this idea, but many other variants exist. In this paper we explore the phenomemology of a particular DM neutrino mass model, containing a triplet scalar. We calculate the relic density and check for constraints from direct detection experiments. The parameter space of the model, allowed by these constraints, contains typically a long-lived or quasi-stable doubly charged scalar, that can be searched for at the LHC. We reinterpret existing searches to derive limits on the masses of the scalars of the model and estimate future sensitivities in the high-luminosity phase of the LHC. The searches we discuss can serve to constrain also many other 1-loop neutrino mass models. | [
{
"version": "v1",
"created": "Tue, 06 Aug 2024 18:00:00 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 20:32:33 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Arbeláez",
"Carolina",
""
],
[
"Cottin",
"Giovanna",
""
],
[
"Helo",
"Juan Carlos",
""
],
[
"Hirsch",
"Martin",
""
],
[
"de Melo",
"Téssio B.",
""
]
] |
2408.04107 | Zeyu Zhang | Zeyu Zhang, Haiying Shen | FDC: Fast KV Dimensionality Compression for Efficient LLM Inference | null | null | null | null | cs.LG cs.DC | http://creativecommons.org/licenses/by/4.0/ | In large-language models, memory constraints in the Key-Value Cache (KVC) pose a challenge during inference. In this work, we propose FDC, a fast KV dimensionality compression system that eliminates the decompression overhead incurred in the existing KV dimensionality compression system, Palu, and reduces attention time. Moreover, FDC employs adaptive compression, tailoring KV compression rates across heads and layers based on their contributions to inference to maximize overall compression while maintaining an accuracy loss constraint. Additionally, FDC enhances the attention kernel to balance the uneven workloads caused by the adaptive compression approach to further reduce attention computation latency. Comprehensive experiments demonstrate that compared to Palu, FDC can reduce Job Completion Time (JCT) by up to 64%, and delivers up to 1.97X throughput under the same latency, while maintaining 99% of the accuracy without compression. When state-of-the-art eviction and quantization methods are combined with FDC, they exhibit similar improvements compared to those combined with Palu. We open-sourced the code. | [
{
"version": "v1",
"created": "Wed, 07 Aug 2024 22:10:26 GMT"
},
{
"version": "v2",
"created": "Wed, 05 Feb 2025 21:44:56 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 20:04:17 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhang",
"Zeyu",
""
],
[
"Shen",
"Haiying",
""
]
] |
2408.04119 | Shohei Wakayama | Shohei Wakayama, Alberto Candela, Paul Hayne, Nisar Ahmed | Active Inference in Contextual Multi-Armed Bandits for Autonomous Robotic Exploration | 11 pages, 12 figures, submitted to IEEE Transactions on Robotics | null | 10.1109/TRO.2025.3577041 | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous selection of optimal options for data collection from multiple alternatives is challenging in uncertain environments. When secondary information about options is accessible, such problems can be framed as contextual multi-armed bandits (CMABs). Neuro-inspired active inference has gained interest for its ability to balance exploration and exploitation using the expected free energy objective function. Unlike previous studies that showed the effectiveness of active inference based strategy for CMABs using synthetic data, this study aims to apply active inference to realistic scenarios, using a simulated mineralogical survey site selection problem. Hyperspectral data from AVIRIS-NG at Cuprite, Nevada, serves as contextual information for predicting outcome probabilities, while geologists' mineral labels represent outcomes. Monte Carlo simulations assess the robustness of active inference against changing expert preferences. Results show that active inference requires fewer iterations than standard bandit approaches with real-world noisy and biased data, and performs better when outcome preferences vary online by adapting the selection strategy to align with expert shifts. | [
{
"version": "v1",
"created": "Wed, 07 Aug 2024 23:00:03 GMT"
},
{
"version": "v2",
"created": "Sun, 05 Jan 2025 17:16:29 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wakayama",
"Shohei",
""
],
[
"Candela",
"Alberto",
""
],
[
"Hayne",
"Paul",
""
],
[
"Ahmed",
"Nisar",
""
]
] |
2408.04287 | Guozhen Shen | Jiaheng Jin and Guozhen Shen | A note on surjective cardinals | 13 pages | null | null | null | math.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For cardinals $\mathfrak{a}$ and $\mathfrak{b}$, we write $\mathfrak{a}=^\ast\mathfrak{b}$ if there are sets $A$ and $B$ of cardinalities $\mathfrak{a}$ and $\mathfrak{b}$, respectively, such that there are partial surjections from $A$ onto $B$ and from $B$ onto $A$. $=^\ast$-equivalence classes are called surjective cardinals. In this article, we show that $\mathsf{ZF}+\mathsf{DC}_\kappa$, where $\kappa$ is a fixed aleph, cannot prove that surjective cardinals form a cardinal algebra, which gives a negative solution to a question proposed by Truss [J. Truss, Ann. Pure Appl. Logic 27, 165--207 (1984)]. Nevertheless, we show that surjective cardinals form a ``surjective cardinal algebra'', whose postulates are almost the same as those of a cardinal algebra, except that the refinement postulate is replaced by the finite refinement postulate. This yields a smoother proof of the cancellation law for surjective cardinals, which states that $m\cdot\mathfrak{a}=^\ast m\cdot\mathfrak{b}$ implies $\mathfrak{a}=^\ast\mathfrak{b}$ for all cardinals $\mathfrak{a},\mathfrak{b}$ and all nonzero natural numbers $m$. | [
{
"version": "v1",
"created": "Thu, 08 Aug 2024 07:53:14 GMT"
},
{
"version": "v2",
"created": "Wed, 14 Aug 2024 06:19:53 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 14:21:17 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Jin",
"Jiaheng",
""
],
[
"Shen",
"Guozhen",
""
]
] |
2408.04802 | Yuta Nozaki | Soichiro Fujii, Yuni Iwamasa, Kei Kimura, Yuta Nozaki, Akira Suzuki | Homotopy types of Hom complexes of graph homomorphisms whose codomains are cycles | 13 pages, 5 figures | null | null | null | math.CO math.AT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For simple graphs $G$ and $H$, the Hom complex $\mathrm{Hom}(G,H)$ is a polyhedral complex whose vertices are the graph homomorphisms $G\to H$ and whose edges connect the pairs of homomorphisms which differ in a single vertex of $G$. Hom complexes play an important role in an algebro-topological approach to the graph coloring problem. It is known that $\mathrm{Hom}(G,H)$ is homotopy equivalent to a disjoint union of points and circles when both $G$ and $H$ are cycles. We generalize this known result by showing that the same holds whenever $G$ is connected and $H$ is a cycle. To this end, we explicitly construct the universal cover of each connected component of $\mathrm{Hom}(G,H)$ and prove that it is contractible. Additionally, we provide a simple criterion to determine whether the connected component containing a given homomorphism is homotopy equivalent to a point or circle. | [
{
"version": "v1",
"created": "Fri, 09 Aug 2024 01:06:21 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 07:55:15 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Fujii",
"Soichiro",
""
],
[
"Iwamasa",
"Yuni",
""
],
[
"Kimura",
"Kei",
""
],
[
"Nozaki",
"Yuta",
""
],
[
"Suzuki",
"Akira",
""
]
] |
2408.04873 | Priyanka Kargupta | Priyanka Kargupta, Yunyi Zhang, Yizhu Jiao, Siru Ouyang, Jiawei Han | Synergizing Unsupervised Episode Detection with LLMs for Large-Scale News Events | ACL 2025 Main Conference. Code available here: https://github.com/pkargupta/epimine | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | State-of-the-art automatic event detection struggles with interpretability and adaptability to evolving large-scale key events -- unlike episodic structures, which excel in these areas. Often overlooked, episodes represent cohesive clusters of core entities performing actions at a specific time and location; a partially ordered sequence of episodes can represent a key event. This paper introduces a novel task, episode detection, which identifies episodes within a news corpus of key event articles. Detecting episodes poses unique challenges, as they lack explicit temporal or locational markers and cannot be merged using semantic similarity alone. While large language models (LLMs) can aid with these reasoning difficulties, they suffer with long contexts typical of news corpora. To address these challenges, we introduce EpiMine, an unsupervised framework that identifies a key event's candidate episodes by leveraging natural episodic partitions in articles, estimated through shifts in discriminative term combinations. These candidate episodes are more cohesive and representative of true episodes, synergizing with LLMs to better interpret and refine them into final episodes. We apply EpiMine to our three diverse, real-world event datasets annotated at the episode level, where it achieves a 59.2% average gain across all metrics compared to baselines. | [
{
"version": "v1",
"created": "Fri, 09 Aug 2024 05:26:31 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 23:08:54 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kargupta",
"Priyanka",
""
],
[
"Zhang",
"Yunyi",
""
],
[
"Jiao",
"Yizhu",
""
],
[
"Ouyang",
"Siru",
""
],
[
"Han",
"Jiawei",
""
]
] |
2408.05177 | Chuwei Wang | Chuwei Wang, Julius Berner, Zongyi Li, Di Zhou, Jiayun Wang, Jane Bae, Anima Anandkumar | Beyond Closure Models: Learning Chaotic-Systems via Physics-Informed Neural Operators | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately predicting the long-term behavior of chaotic systems is crucial for various applications such as climate modeling. However, achieving such predictions typically requires iterative computations over a dense spatiotemporal grid to account for the unstable nature of chaotic systems, which is expensive and impractical in many real-world situations. An alternative approach to such a full-resolved simulation is using a coarse grid and then correcting its errors through a \textit{closure model}, which approximates the overall information from fine scales not captured in the coarse-grid simulation. Recently, ML approaches have been used for closure modeling, but they typically require a large number of training samples from expensive fully-resolved simulations (FRS). In this work, we prove an even more fundamental limitation, i.e., the standard approach to learning closure models suffers from a large approximation error for generic problems, no matter how large the model is, and it stems from the non-uniqueness of the mapping. We propose an alternative end-to-end learning approach using a physics-informed neural operator (PINO) that overcomes this limitation by not using a closure model or a coarse-grid solver. We first train the PINO model on data from a coarse-grid solver and then fine-tune it with (a small amount of) FRS and physics-based losses on a fine grid. The discretization-free nature of neural operators means that they do not suffer from the restriction of a coarse grid that closure models face, and they can provably approximate the long-term statistics of chaotic systems. In our experiments, our PINO model achieves a 330x speedup compared to FRS with a relative error $\sim 10\%$. In contrast, the closure model coupled with a coarse-grid solver is $60$x slower than PINO while having a much higher error $\sim186\%$ when the closure model is trained on the same FRS dataset. | [
{
"version": "v1",
"created": "Fri, 09 Aug 2024 17:05:45 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Oct 2024 07:58:34 GMT"
},
{
"version": "v3",
"created": "Thu, 10 Oct 2024 03:54:45 GMT"
},
{
"version": "v4",
"created": "Mon, 09 Jun 2025 07:37:52 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wang",
"Chuwei",
""
],
[
"Berner",
"Julius",
""
],
[
"Li",
"Zongyi",
""
],
[
"Zhou",
"Di",
""
],
[
"Wang",
"Jiayun",
""
],
[
"Bae",
"Jane",
""
],
[
"Anandkumar",
"Anima",
""
]
] |
2408.05762 | Andrew Ryzhikov | Stefan Kiefer, Andrew Ryzhikov | The complexity of computing the period and the exponent of a digraph | minor edits | null | 10.1016/j.ipl.2025.106590 | null | cs.DM math.CO | http://creativecommons.org/licenses/by/4.0/ | The period of a strongly connected digraph is the greatest common divisor of the lengths of all its cycles. The period of a digraph is the least common multiple of the periods of its strongly connected components. These notions play an important role in the theory of Markov chains and the analysis of powers of nonnegative matrices. While the time complexity of computing the period is well-understood, little is known about its space complexity. We show that the problem of computing the period of a digraph is NL-complete, even if all its cycles are contained in the same strongly connected component. However, if the digraph is strongly connected, we show that this problem becomes L-complete. For primitive digraphs (that is, strongly connected digraphs of period one), there always exists a number $m$ such that there is a path of length exactly $m$ between every two vertices. We show that computing the smallest such $m$, called the exponent of a digraph, is NL-complete. The exponent of a primitive digraph is a particular case of the index of convergence of a nonnegative matrix, which we also show to be computable in NL, and thus NL-complete. | [
{
"version": "v1",
"created": "Sun, 11 Aug 2024 12:48:46 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 13:16:35 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kiefer",
"Stefan",
""
],
[
"Ryzhikov",
"Andrew",
""
]
] |
2408.06616 | Jeongseok Oh | Indranil Biswas, Nilkantha Das, Jeongseok Oh, Anantadulal Paul | Gromov-Witten invariants in family and quantum cohomology | null | null | null | null | math.AG | http://creativecommons.org/licenses/by/4.0/ | A moduli space of stable maps to the fibers of a fiber bundle is constructed. The new moduli space is a family version of the classical moduli space of stable maps to a non-singular complex projective variety. The virtual cycle for this moduli space is also constructed, and an analogue of Gromov-Witten invariants is defined. As an application, we recover the formula for the number of rational degree d curves in P3, whose image lies in a plane in P3 (known as planar curves in P3), intersecting r general lines while passing through given s general points, where r + 2s = 3d + 2, firstly proved by R. Mukherjee, R. Kumar Singh and the fourth named author. | [
{
"version": "v1",
"created": "Tue, 13 Aug 2024 04:06:13 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 01:14:17 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Biswas",
"Indranil",
""
],
[
"Das",
"Nilkantha",
""
],
[
"Oh",
"Jeongseok",
""
],
[
"Paul",
"Anantadulal",
""
]
] |
2408.07036 | Karol Urba\'nski | Karol Urba\'nski | Visual relativistic mechanics | 12 pages, 11 figures | null | null | null | gr-qc physics.class-ph physics.ed-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article shows how to express relativistic concepts in a visual manner using the full power of hyperbolic trigonometric functions. Minkowski diagrams in energy-momentum space are used in conjunction with hyperbolic triangles. Elegant new derivations of the relativistic rocket equation and the relativistic Doppler effect are presented that use this visual approach. | [
{
"version": "v1",
"created": "Tue, 13 Aug 2024 17:02:20 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 22:19:55 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Urbański",
"Karol",
""
]
] |
2408.07116 | Sean J. Liu | Sean J. Liu, Nupur Kumari, Ariel Shamir, Jun-Yan Zhu | Generative Photomontage | CVPR 2025. Project webpage: https://lseancs.github.io/generativephotomontage/ | In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2025 | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-to-image models are powerful tools for image creation. However, the generation process is akin to a dice roll and makes it difficult to achieve a single image that captures everything a user wants. In this paper, we propose a framework for creating the desired image by compositing it from various parts of generated images, in essence forming a Generative Photomontage. Given a stack of images generated by ControlNet using the same input condition and different seeds, we let users select desired parts from the generated results using a brush stroke interface. We introduce a novel technique that takes in the user's brush strokes, segments the generated images using a graph-based optimization in diffusion feature space, and then composites the segmented regions via a new feature-space blending method. Our method faithfully preserves the user-selected regions while compositing them harmoniously. We demonstrate that our flexible framework can be used for many applications, including generating new appearance combinations, fixing incorrect shapes and artifacts, and improving prompt alignment. We show compelling results for each application and demonstrate that our method outperforms existing image blending methods and various baselines. | [
{
"version": "v1",
"created": "Tue, 13 Aug 2024 17:59:51 GMT"
},
{
"version": "v2",
"created": "Fri, 16 Aug 2024 17:59:51 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 03:36:08 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Liu",
"Sean J.",
""
],
[
"Kumari",
"Nupur",
""
],
[
"Shamir",
"Ariel",
""
],
[
"Zhu",
"Jun-Yan",
""
]
] |
2408.07352 | Boris Betancourt Kamenetskaia | Boris Betancourt Kamenetskaia, Nissim Fraija and Gonzalo Herrera | Polarization Measurements as a Probe of Axion-Photon Coupling: a Study of GRB 221009A | 10 pages, 4 figures | null | 10.1103/2fmw-zz63 | null | hep-ph astro-ph.CO astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Axion Like Particles (ALPs) can be produced in Gamma Ray Bursts, altering the polarization of the electromagnetic emission in these events. For the first time, we derive bounds on the axion-photon coupling from polarization measurements of GRB 221009A, performing a full calculation of the Stokes parameters, as it is typically done in the astrophysics community. Within astrophysical uncertainties, our limits on the axion-photon coupling are competitive with complementary probes in the axion mass range $10^{-9}$ eV $\lesssim m_a \lesssim 10^{-8}$ eV, further allowing to probe motivated parameter space of ALP dark matter. | [
{
"version": "v1",
"created": "Wed, 14 Aug 2024 07:52:16 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 18:23:35 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kamenetskaia",
"Boris Betancourt",
""
],
[
"Fraija",
"Nissim",
""
],
[
"Herrera",
"Gonzalo",
""
]
] |
2408.08208 | Bohao Wang | Bohao Wang, Feng Liu, Changwang Zhang, Jiawei Chen, Yudi Wu, Sheng Zhou, Xingyu Lou, Jun Wang, Yan Feng, Chun Chen, Can Wang | LLM4DSR: Leveraging Large Language Model for Denoising Sequential Recommendation | null | null | null | null | cs.IR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sequential Recommenders generate recommendations based on users' historical interaction sequences. However, in practice, these collected sequences are often contaminated by noisy interactions, which significantly impairs recommendation performance. Accurately identifying such noisy interactions without additional information is particularly challenging due to the absence of explicit supervisory signals indicating noise. Large Language Models (LLMs), equipped with extensive open knowledge and semantic reasoning abilities, offer a promising avenue to bridge this information gap. However, employing LLMs for denoising in sequential recommendation presents notable challenges: 1) Direct application of pretrained LLMs may not be competent for the denoising task, frequently generating nonsensical responses; 2) Even after fine-tuning, the reliability of LLM outputs remains questionable, especially given the complexity of the denoising task and the inherent hallucinatory issue of LLMs.
To tackle these challenges, we propose LLM4DSR, a tailored approach for denoising sequential recommendation using LLMs. We constructed a self-supervised fine-tuning task to activate LLMs' capabilities to identify noisy items and suggest replacements. Furthermore, we developed an uncertainty estimation module that ensures only high-confidence responses are utilized for sequence corrections. Remarkably, LLM4DSR is model-agnostic, allowing corrected sequences to be flexibly applied across various recommendation models. Extensive experiments validate the superiority of LLM4DSR over existing methods. | [
{
"version": "v1",
"created": "Thu, 15 Aug 2024 15:18:46 GMT"
},
{
"version": "v2",
"created": "Tue, 26 Nov 2024 08:07:08 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 04:28:59 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wang",
"Bohao",
""
],
[
"Liu",
"Feng",
""
],
[
"Zhang",
"Changwang",
""
],
[
"Chen",
"Jiawei",
""
],
[
"Wu",
"Yudi",
""
],
[
"Zhou",
"Sheng",
""
],
[
"Lou",
"Xingyu",
""
],
[
"Wang",
"Jun",
""
],
[
"Feng",
"Yan",
""
],
[
"Chen",
"Chun",
""
],
[
"Wang",
"Can",
""
]
] |
2408.08968 | Cyril Shih-Huan Hsu | Cyril Shih-Huan Hsu, Danny De Vleeschauwer, Chrysa Papagianni, Paola Grosso | Online SLA Decomposition: Enabling Real-Time Adaptation to Evolving Network Systems | The paper has been accepted for publication at EuCNC & 6G Summit 2025 | null | null | null | cs.NI cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When a network slice spans multiple technology domains, it is crucial for each domain to uphold the End-to-End (E2E) Service Level Agreement (SLA) associated with the slice. Consequently, the E2E SLA must be properly decomposed into partial SLAs that are assigned to each domain involved. In a network slice management system with a two-level architecture, comprising an E2E service orchestrator and local domain controllers, we consider that the orchestrator has access only to historical data regarding the responses of local controllers to previous requests, and this information is used to construct a risk model for each domain. In this study, we extend our previous work by investigating the dynamic nature of real-world systems and introducing an online learning-decomposition framework to tackle the dynamicity. We propose a framework that continuously updates the risk models based on the most recent feedback. This approach leverages key components such as online gradient descent and FIFO memory buffers, which enhance the stability and robustness of the overall process. Our empirical study on an analytic model-based simulator demonstrates that the proposed framework outperforms the state-of-the-art static approach, delivering more accurate and resilient SLA decomposition under varying conditions and data limitations. Furthermore, we provide a comprehensive complexity analysis of the proposed solution. | [
{
"version": "v1",
"created": "Fri, 16 Aug 2024 18:34:11 GMT"
},
{
"version": "v2",
"created": "Tue, 20 Aug 2024 11:17:56 GMT"
},
{
"version": "v3",
"created": "Thu, 05 Dec 2024 11:01:30 GMT"
},
{
"version": "v4",
"created": "Fri, 11 Apr 2025 16:19:31 GMT"
},
{
"version": "v5",
"created": "Sat, 07 Jun 2025 08:59:14 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Hsu",
"Cyril Shih-Huan",
""
],
[
"De Vleeschauwer",
"Danny",
""
],
[
"Papagianni",
"Chrysa",
""
],
[
"Grosso",
"Paola",
""
]
] |
2408.09121 | Yuan Tian | Yuan Tian, Tianyi Zhang | Selective Prompt Anchoring for Code Generation | Accepted by ICML'25 | null | null | null | cs.LG cs.AI cs.CL cs.SE | http://creativecommons.org/licenses/by/4.0/ | Recent advances in large language models (LLMs) have transformed software development by automatically generating code from natural language. Yet challenges remain in generating fully correct code that aligns with user intent. Our study reveals that LLMs tend to pay less attention to user prompts as more code tokens are generated. We hypothesize that this attention dilution issue is an important reason for code generation errors. To mitigate this issue, we propose Selective Prompt Anchoring (SPA) to guide code LLMs to pay more attention to user intent when generating code. We evaluate SPA using six base LLMs across six benchmarks. Our results demonstrate that SPA enhances Pass@1 by up to 12.9%, consistently outperforming SOTA code generation methods in all settings. Our code is available at https://github.com/magic-YuanTian/Selective-Prompt-Anchoring. | [
{
"version": "v1",
"created": "Sat, 17 Aug 2024 07:11:02 GMT"
},
{
"version": "v2",
"created": "Wed, 21 Aug 2024 06:01:08 GMT"
},
{
"version": "v3",
"created": "Thu, 06 Feb 2025 23:16:06 GMT"
},
{
"version": "v4",
"created": "Fri, 21 Feb 2025 03:02:41 GMT"
},
{
"version": "v5",
"created": "Sun, 08 Jun 2025 20:42:38 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Tian",
"Yuan",
""
],
[
"Zhang",
"Tianyi",
""
]
] |
2408.09568 | Meghdad Dehghan | Meghdad Dehghan, Jie JW Wu, Fatemeh H. Fard and Ali Ouni | MergeRepair: An Exploratory Study on Merging Task-Specific Adapters in Code LLMs for Automated Program Repair | null | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) have shown high capabilities in several software development-related tasks such as program repair, documentation, code refactoring, debugging, and testing. However, training these models requires massive amount of data and significant computational resources. Adapters are specialized, small modules designed for parameter efficient fine-tuning of LLMs for specific tasks, domains, or applications without requiring extensive retraining of the entire model. These adapters offer a more efficient way to customize LLMs for particular needs, leveraging the pre-existing capabilities of the large model. Model (and adapter) merging have emerged as a technique to develop one model capable of multiple tasks, with minimal or no training required. Although model and adapter merging has shown promising performance in domains such as natural language processing and computer vision, its applicability to software engineering tasks remains underexplored. In this paper, we investigate the effectiveness of merged adapters within the context of software engineering, with a particular focus on the Automated Program Repair (APR) task, through our approach, MergeRepair. In particular, we merge multiple task-specific adapters using three different merging methods, including weight-averaging, ties, and dare-ties, and evaluate the performance of the merged adapter on the APR task. We introduce a continual merging approach, a novel method in which we sequentially merge the task-specific adapters where the order and weight of the merged adapters play a significant role. We further compare the performance of our approach with a baseline method consisting of equal-weight merging applied on parameters of different adapters, where all adapters are of equal importance. | [
{
"version": "v1",
"created": "Sun, 18 Aug 2024 18:45:48 GMT"
},
{
"version": "v2",
"created": "Mon, 26 Aug 2024 19:27:46 GMT"
},
{
"version": "v3",
"created": "Fri, 06 Jun 2025 21:09:35 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Dehghan",
"Meghdad",
""
],
[
"Wu",
"Jie JW",
""
],
[
"Fard",
"Fatemeh H.",
""
],
[
"Ouni",
"Ali",
""
]
] |
2408.09820 | Boris Volkov | Ivan Russkikh, Boris Volkov, Alexander Pechen | Quantum channels, complex Stiefel manifolds, and optimization | 21 pages | Quantum Information Processing, 24 (2025) 167 | 10.1007/s11128-025-04782-x | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Most general dynamics of an open quantum system is commonly represented by a quantum channel, which is a completely positive trace-preserving map (CPTP or Kraus map). Well-known are the representations of quantum channels by Choi matrices and by Kraus operator-sum representation (OSR). As was shown before, one can use Kraus OSR to parameterize quantum channels by points of a suitable quotient under the action of the unitary group of some complex Stiefel manifold. In this work, we establish a continuity relation (homeomorphism) between the topological space of quantum channels and the quotient of the complex Stiefel manifold. Then the metric on the set of quantum channels induced by the Riemannian metric on the Stiefel manifold is defined. The established relation can be applied to various quantum optimization problems. As an example, we apply it to the analysis of extrema points for a wide variety of quantum control objective functionals defined on the complex Stiefel manifolds, including mean value, generation of quantum gates, thermodynamic quantities involving entropy, etc. | [
{
"version": "v1",
"created": "Mon, 19 Aug 2024 09:15:54 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Russkikh",
"Ivan",
""
],
[
"Volkov",
"Boris",
""
],
[
"Pechen",
"Alexander",
""
]
] |
2408.10903 | Yeyong Yu | Yeyong Yu, Runsheng Yu, Haojie Wei, Zhanqiu Zhang, Quan Qian | BEYOND DIALOGUE: A Profile-Dialogue Alignment Framework Towards General Role-Playing Language Model | null | null | null | null | cs.CL cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid advancement of large language models (LLMs) has revolutionized role-playing, enabling the development of general role-playing models. However, current role-playing training has two significant issues: (I) Using a predefined role profile to prompt dialogue training for specific scenarios usually leads to inconsistencies and even conflicts between the dialogue and the profile, resulting in training biases. (II) The model learns to imitate the role based solely on the profile, neglecting profile-dialogue alignment at the sentence level. In this work, we propose a simple yet effective framework called BEYOND DIALOGUE, designed to overcome these hurdles. This framework innovatively introduces "beyond dialogue" tasks to align dialogue with profile traits based on each specific scenario, thereby eliminating biases during training. Furthermore, by adopting an innovative prompting mechanism that generates reasoning outcomes for training, the framework allows the model to achieve fine-grained alignment between profile and dialogue at the sentence level. The aforementioned methods are fully automated and low-cost. Additionally, the integration of automated dialogue and objective evaluation methods forms a comprehensive framework, paving the way for general role-playing. Experimental results demonstrate that our model excels in adhering to and reflecting various dimensions of role profiles, outperforming most proprietary general and specialized role-playing baselines. All code and datasets are available at https://github.com/yuyouyu32/BeyondDialogue. | [
{
"version": "v1",
"created": "Tue, 20 Aug 2024 14:47:38 GMT"
},
{
"version": "v2",
"created": "Wed, 21 Aug 2024 03:31:25 GMT"
},
{
"version": "v3",
"created": "Mon, 26 Aug 2024 07:37:19 GMT"
},
{
"version": "v4",
"created": "Tue, 27 Aug 2024 02:58:39 GMT"
},
{
"version": "v5",
"created": "Thu, 29 Aug 2024 02:38:05 GMT"
},
{
"version": "v6",
"created": "Mon, 09 Jun 2025 06:26:33 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yu",
"Yeyong",
""
],
[
"Yu",
"Runsheng",
""
],
[
"Wei",
"Haojie",
""
],
[
"Zhang",
"Zhanqiu",
""
],
[
"Qian",
"Quan",
""
]
] |
2408.11222 | Jacob Z. Shapiro | Andr\'es Larra\'in-Hubach and Jacob Shapiro | Semiclassical estimates for the magnetic Schr\"odinger operator on the line | 28 pages | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | We prove a weighted Carleman estimate for a class of one-dimensional, self-adjoint Schr\"odinger operators $P(h)$ with low regularity electric and magnetic potentials, where $h > 0$ is a semiclassical parameter. The long range part of either potential has bounded variation. The short range part of the magnetic potential belongs to $L^1(\mathbb{R}) \cap L^2(\mathbb{R})$, while the short range part of the electric potential is a finite signed measure. The proof is a one dimensional instance of the energy method, which is used to prove Carleman estimates in higher dimensions and in more complicated geometries. The novelty of our result lies in the weak regularity assumptions on the coefficients. As a consequence of the Carleman estimate, we establish an optimal limiting absorption resolvent estimates for $P(h)$. We also present standard applications to the distribution of resonances for $P(1)$ and to associated evolution equations. | [
{
"version": "v1",
"created": "Tue, 20 Aug 2024 22:40:38 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 09:27:05 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Larraín-Hubach",
"Andrés",
""
],
[
"Shapiro",
"Jacob",
""
]
] |
2408.11317 | Hongfa Zi | Hongfa Zi, Zhen Liu | The continuous accumulation of civilization core in the cycle of elements-creature, benefits and weapons | 14 pages, 7 figures | null | null | null | cs.OH | http://creativecommons.org/licenses/by/4.0/ | The comprehensive strength of a country varies from strong to weak, divided into three condition: descending, periodicity destruction or rapidly rising, Exploring the differences can solve the development crisis. the most important things for a country are interests, weapons and creature, corresponding to money, technology and people. The ship industry has two attribute of financial benefits and technological weapons. Commercial ships can transport massive commodity and warships carry updating of massive technological weapons; But a new core: equity incentives have emerged, and it has helped the rapid development of the computer industry. This article uses comparative analysis and comparative historical analysis to observe the changes in the United States and China after the mutual circulation of two elements and the double circulation of three elements in history, such as the growth rates of GDP and patent applications. Then, it summarizes the changes brought by the core of civilization to the country.Through this article, it can be concluded that the core of civilization consists of ships, equity incentives, record-wisdom method; Through the circulation of new elements, a country can transform into civilizations with three cycles, achieving mutual circulation among the three and enhancing endogenous power; The core of civilization can enhance the stability of economic development, prevent economic crises, and achieve a more balanced civilization. | [
{
"version": "v1",
"created": "Wed, 21 Aug 2024 03:51:40 GMT"
},
{
"version": "v2",
"created": "Fri, 30 Aug 2024 00:31:05 GMT"
},
{
"version": "v3",
"created": "Wed, 09 Oct 2024 14:29:22 GMT"
},
{
"version": "v4",
"created": "Thu, 17 Oct 2024 19:32:24 GMT"
},
{
"version": "v5",
"created": "Tue, 05 Nov 2024 13:27:27 GMT"
},
{
"version": "v6",
"created": "Thu, 20 Mar 2025 07:44:03 GMT"
},
{
"version": "v7",
"created": "Sun, 08 Jun 2025 14:41:50 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zi",
"Hongfa",
""
],
[
"Liu",
"Zhen",
""
]
] |
2408.11600 | Renlong Wang | Shutian Cui, Renlong Wang | A Novel {\delta}-SBM-OPA Approach for Policy-Driven Analysis of Carbon Emission Efficiency under Uncertainty in the Chinese Industrial Sector | N.A | null | 10.1016/j.energy.2025.135832 | null | econ.GN q-fin.EC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Regional differences in carbon emission efficiency arise from disparities in resource distribution, industrial structure, and development level, which are often influenced by government policy preferences. However, currently, most studies fail to consider the impact of government policy preferences and data uncertainty on carbon emission efficiency. To address the above limitations, this study proposes a hybrid model based on $\delta$-slack-based model ($\delta$-SBM) and ordinal priority approach (OPA) for measuring carbon emission efficiency driven by government policy preferences under data uncertainty. The proposed $\delta$-SBM-OPA model incorporates constraints on the importance of input and output variables under different policy preference scenarios. It then develops the efficiency optimization model with Farrell frontiers and efficiency tapes to deal with the data uncertainty in input and output variables. This study demonstrates the proposed model by analyzing industrial carbon emission efficiency of Chinese provinces in 2021. It examines the carbon emission efficiency and corresponding clustering results of provinces under three types of policies: economic priority, environmental priority, and technological priority, with varying priority preferences. The results indicate that the carbon emission efficiency of the 30 provinces can mainly be categorized into technology-driven, development-balanced, and transition-potential types, with most provinces achieving optimal efficiency under the technology-dominant preferences across all policy scenarios. Ultimately, this study suggests a tailored roadmap and crucial initiatives for different provinces to progressively and systematically work towards achieving the low carbon goal. | [
{
"version": "v1",
"created": "Wed, 21 Aug 2024 13:13:27 GMT"
},
{
"version": "v2",
"created": "Wed, 18 Dec 2024 15:20:09 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Cui",
"Shutian",
""
],
[
"Wang",
"Renlong",
""
]
] |
2408.11800 | Rounak Meyur | Rounak Meyur, Hung Phan, Sridevi Wagle, Jan Strube, Mahantesh Halappanavar, Sameera Horawalavithana, Anurag Acharya, Sai Munikoti | WeQA: A Benchmark for Retrieval Augmented Generation in Wind Energy Domain | 8 pages without Limitation and References | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Wind energy project assessments present significant challenges for decision-makers, who must navigate and synthesize hundreds of pages of environmental and scientific documentation. These documents often span different regions and project scales, covering multiple domains of expertise. This process traditionally demands immense time and specialized knowledge from decision-makers. The advent of Large Language Models (LLM) and Retrieval Augmented Generation (RAG) approaches offer a transformative solution, enabling rapid, accurate cross-document information retrieval and synthesis. As the landscape of Natural Language Processing (NLP) and text generation continues to evolve, benchmarking becomes essential to evaluate and compare the performance of different RAG-based LLMs. In this paper, we present a comprehensive framework to generate a domain relevant RAG benchmark. Our framework is based on automatic question-answer generation with Human (domain experts)-AI (LLM) teaming. As a case study, we demonstrate the framework by introducing WeQA, a first-of-its-kind benchmark on the wind energy domain which comprises of multiple scientific documents/reports related to environmental aspects of wind energy projects. Our framework systematically evaluates RAG performance using diverse metrics and multiple question types with varying complexity level, providing a foundation for rigorous assessment of RAG-based systems in complex scientific domains and enabling researchers to identify areas for improvement in domain-specific applications. | [
{
"version": "v1",
"created": "Wed, 21 Aug 2024 17:43:11 GMT"
},
{
"version": "v2",
"created": "Tue, 24 Sep 2024 22:50:15 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 14:33:22 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Meyur",
"Rounak",
""
],
[
"Phan",
"Hung",
""
],
[
"Wagle",
"Sridevi",
""
],
[
"Strube",
"Jan",
""
],
[
"Halappanavar",
"Mahantesh",
""
],
[
"Horawalavithana",
"Sameera",
""
],
[
"Acharya",
"Anurag",
""
],
[
"Munikoti",
"Sai",
""
]
] |
2408.11878 | Xiao-Yang Liu | Jimin Huang, Mengxi Xiao, Dong Li, Zihao Jiang, Yuzhe Yang, Yifei Zhang, Lingfei Qian, Yan Wang, Xueqing Peng, Yang Ren, Ruoyu Xiang, Zhengyu Chen, Xiao Zhang, Yueru He, Weiguang Han, Shunian Chen, Lihang Shen, Daniel Kim, Yangyang Yu, Yupeng Cao, Zhiyang Deng, Haohang Li, Duanyu Feng, Yongfu Dai, VijayaSai Somasundaram, Peng Lu, Guojun Xiong, Zhiwei Liu, Zheheng Luo, Zhiyuan Yao, Ruey-Ling Weng, Meikang Qiu, Kaleb E Smith, Honghai Yu, Yanzhao Lai, Min Peng, Jian-Yun Nie, Jordan W. Suchow, Xiao-Yang Liu, Benyou Wang, Alejandro Lopez-Lira, Qianqian Xie, Sophia Ananiadou and Junichi Tsujii | Open-FinLLMs: Open Multimodal Large Language Models for Financial Applications | 33 pages, 13 figures | null | null | null | cs.CL cs.CE q-fin.CP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Financial LLMs hold promise for advancing financial tasks and domain-specific applications. However, they are limited by scarce corpora, weak multimodal capabilities, and narrow evaluations, making them less suited for real-world application. To address this, we introduce \textit{Open-FinLLMs}, the first open-source multimodal financial LLMs designed to handle diverse tasks across text, tabular, time-series, and chart data, excelling in zero-shot, few-shot, and fine-tuning settings. The suite includes FinLLaMA, pre-trained on a comprehensive 52-billion-token corpus; FinLLaMA-Instruct, fine-tuned with 573K financial instructions; and FinLLaVA, enhanced with 1.43M multimodal tuning pairs for strong cross-modal reasoning. We comprehensively evaluate Open-FinLLMs across 14 financial tasks, 30 datasets, and 4 multimodal tasks in zero-shot, few-shot, and supervised fine-tuning settings, introducing two new multimodal evaluation datasets. Our results show that Open-FinLLMs outperforms afvanced financial and general LLMs such as GPT-4, across financial NLP, decision-making, and multi-modal tasks, highlighting their potential to tackle real-world challenges. To foster innovation and collaboration across academia and industry, we release all codes (https://anonymous.4open.science/r/PIXIU2-0D70/B1D7/LICENSE) and models under OSI-approved licenses. | [
{
"version": "v1",
"created": "Tue, 20 Aug 2024 16:15:28 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Apr 2025 14:18:35 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 01:25:35 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Huang",
"Jimin",
""
],
[
"Xiao",
"Mengxi",
""
],
[
"Li",
"Dong",
""
],
[
"Jiang",
"Zihao",
""
],
[
"Yang",
"Yuzhe",
""
],
[
"Zhang",
"Yifei",
""
],
[
"Qian",
"Lingfei",
""
],
[
"Wang",
"Yan",
""
],
[
"Peng",
"Xueqing",
""
],
[
"Ren",
"Yang",
""
],
[
"Xiang",
"Ruoyu",
""
],
[
"Chen",
"Zhengyu",
""
],
[
"Zhang",
"Xiao",
""
],
[
"He",
"Yueru",
""
],
[
"Han",
"Weiguang",
""
],
[
"Chen",
"Shunian",
""
],
[
"Shen",
"Lihang",
""
],
[
"Kim",
"Daniel",
""
],
[
"Yu",
"Yangyang",
""
],
[
"Cao",
"Yupeng",
""
],
[
"Deng",
"Zhiyang",
""
],
[
"Li",
"Haohang",
""
],
[
"Feng",
"Duanyu",
""
],
[
"Dai",
"Yongfu",
""
],
[
"Somasundaram",
"VijayaSai",
""
],
[
"Lu",
"Peng",
""
],
[
"Xiong",
"Guojun",
""
],
[
"Liu",
"Zhiwei",
""
],
[
"Luo",
"Zheheng",
""
],
[
"Yao",
"Zhiyuan",
""
],
[
"Weng",
"Ruey-Ling",
""
],
[
"Qiu",
"Meikang",
""
],
[
"Smith",
"Kaleb E",
""
],
[
"Yu",
"Honghai",
""
],
[
"Lai",
"Yanzhao",
""
],
[
"Peng",
"Min",
""
],
[
"Nie",
"Jian-Yun",
""
],
[
"Suchow",
"Jordan W.",
""
],
[
"Liu",
"Xiao-Yang",
""
],
[
"Wang",
"Benyou",
""
],
[
"Lopez-Lira",
"Alejandro",
""
],
[
"Xie",
"Qianqian",
""
],
[
"Ananiadou",
"Sophia",
""
],
[
"Tsujii",
"Junichi",
""
]
] |
2408.11884 | Jingying Ma | Jingying Ma, Qika Lin, Ziyu Jia, Mengling Feng | ST-USleepNet: A Spatial-Temporal Coupling Prominence Network for Multi-Channel Sleep Staging | null | null | null | null | q-bio.NC cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sleep staging is critical to assess sleep quality and diagnose disorders. Despite advancements in artificial intelligence enabling automated sleep staging, significant challenges remain: (1) Simultaneously extracting prominent temporal and spatial sleep features from multi-channel raw signals, including characteristic sleep waveforms and salient spatial brain networks. (2) Capturing the spatial-temporal coupling patterns essential for accurate sleep staging. To address these challenges, we propose a novel framework named ST-USleepNet, comprising a spatial-temporal graph construction module (ST) and a U-shaped sleep network (USleepNet). The ST module converts raw signals into a spatial-temporal graph based on signal similarity, temporal, and spatial relationships to model spatial-temporal coupling patterns. The USleepNet employs a U-shaped structure for both the temporal and spatial streams, mirroring its original use in image segmentation to isolate significant targets. Applied to raw sleep signals and graph data from the ST module, USleepNet effectively segments these inputs, simultaneously extracting prominent temporal and spatial sleep features. Testing on three datasets demonstrates that ST-USleepNet outperforms existing baselines, and model visualizations confirm its efficacy in extracting prominent sleep features and temporal-spatial coupling patterns across various sleep stages. The code is available at https://github.com/Majy-Yuji/ST-USleepNet. | [
{
"version": "v1",
"created": "Wed, 21 Aug 2024 14:57:44 GMT"
},
{
"version": "v2",
"created": "Thu, 23 Jan 2025 03:45:29 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 15:25:02 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Ma",
"Jingying",
""
],
[
"Lin",
"Qika",
""
],
[
"Jia",
"Ziyu",
""
],
[
"Feng",
"Mengling",
""
]
] |
2408.12206 | Yuki Mifune | Souvik Dey, Yuki Mifune | Upper bounds for dimensions of singularity categories and their annihilators | 19 pages, coauthor added in this version | null | null | null | math.AC math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $R$ be a commutative noetherian ring. Denote by $\operatorname{mod} R$ the category of finitely generated $R$-modules and by $\operatorname{D^b}(R)$ the bounded derived category of $\operatorname{mod} R$. In this paper, we first investigate localizations and annihilators of Verdier quotients of $\operatorname{D^b}(R)$. After that, we explore upper bounds for the dimension of the singularity category of $R$ and its (strong) generators. We extend a theorem of Liu to the case where $R$ is neither an isolated singularity nor even a local ring. Some of our results are more generally stated in terms of Spanier--Whitehead category of a resolving subcategory. | [
{
"version": "v1",
"created": "Thu, 22 Aug 2024 08:30:42 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 13:32:12 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Dey",
"Souvik",
""
],
[
"Mifune",
"Yuki",
""
]
] |
2408.12483 | Shaobo Wang | Shaobo Wang, Yantai Yang, Qilong Wang, Kaixin Li, Linfeng Zhang, and Junchi Yan | Not All Samples Should Be Utilized Equally: Towards Understanding and Improving Dataset Distillation | Accepted by Synthetic Data for Computer Vision Workshop at CVPR, 2025 | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Dataset Distillation (DD) aims to synthesize a small dataset capable of performing comparably to the original dataset. Despite the success of numerous DD methods, theoretical exploration of this area remains unaddressed. In this paper, we take an initial step towards understanding various matching-based DD methods from the perspective of sample difficulty. We begin by empirically examining sample difficulty, measured by gradient norm, and observe that different matching-based methods roughly correspond to specific difficulty tendencies. We then extend the neural scaling laws of data pruning to DD to theoretically explain these matching-based methods. Our findings suggest that prioritizing the synthesis of easier samples from the original dataset can enhance the quality of distilled datasets, especially in low IPC (image-per-class) settings. Based on our empirical observations and theoretical analysis, we introduce the Sample Difficulty Correction (SDC) approach, designed to predominantly generate easier samples to achieve higher dataset quality. Our SDC can be seamlessly integrated into existing methods as a plugin with minimal code adjustments. Experimental results demonstrate that adding SDC generates higher-quality distilled datasets across 7 distillation methods and 6 datasets. | [
{
"version": "v1",
"created": "Thu, 22 Aug 2024 15:20:32 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 07:55:16 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wang",
"Shaobo",
""
],
[
"Yang",
"Yantai",
""
],
[
"Wang",
"Qilong",
""
],
[
"Li",
"Kaixin",
""
],
[
"Zhang",
"Linfeng",
""
],
[
"Yan",
"Junchi",
""
]
] |
2408.12766 | Siddhant Gautam | Siddhant Gautam, Marc L. Klasky, Balasubramanya T. Nadiga, Trevor Wilcox, Gary Salazar, and Saiprasad Ravishankar | Learning Robust Features for Scatter Removal and Reconstruction in Dynamic ICF X-Ray Tomography | null | null | null | null | eess.IV | http://creativecommons.org/licenses/by/4.0/ | Density reconstruction from X-ray projections is an important problem in radiography with key applications in scientific and industrial X-ray computed tomography (CT). Often, such projections are corrupted by unknown sources of noise and scatter, which when not properly accounted for, can lead to significant errors in density reconstruction. In the setting of this problem, recent deep learning-based methods have shown promise in improving the accuracy of density reconstruction. In this article, we propose a deep learning-based encoder-decoder framework wherein the encoder extracts robust features from noisy/corrupted X-ray projections and the decoder reconstructs the density field from the features extracted by the encoder. We explore three options for the latent-space representation of features: physics-inspired supervision, self-supervision, and no supervision. We find that variants based on self-supervised and physicsinspired supervised features perform better over a range of unknown scatter and noise. In extreme noise settings, the variant with self-supervised features performs best. After investigating further details of the proposed deep-learning methods, we conclude by demonstrating that the newly proposed methods are able to achieve higher accuracy in density reconstruction when compared to a traditional iterative technique. | [
{
"version": "v1",
"created": "Thu, 22 Aug 2024 23:42:34 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 06:03:58 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Gautam",
"Siddhant",
""
],
[
"Klasky",
"Marc L.",
""
],
[
"Nadiga",
"Balasubramanya T.",
""
],
[
"Wilcox",
"Trevor",
""
],
[
"Salazar",
"Gary",
""
],
[
"Ravishankar",
"Saiprasad",
""
]
] |
2408.13147 | Yihao Zhang | Yihao Zhang, Harpreet S. Sawhney, John J. Leonard | ShapeICP: Iterative Category-level Object Pose and Shape Estimation from Depth | null | null | null | null | cs.CV cs.AI cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Category-level object pose and shape estimation from a single depth image has recently drawn research attention due to its potential utility for tasks such as robotics manipulation. The task is particularly challenging because the three unknowns, object pose, object shape, and model-to-measurement correspondences, are compounded together, but only a single view of depth measurements is provided. Most of the prior work heavily relies on data-driven approaches to obtain solutions to at least one of the unknowns, and typically two, running with the risk of failing to generalize to unseen domains. The shape representations used in the prior work also mainly focus on point cloud and signed distance field (SDF). In stark contrast to the prior work, we approach the problem using an iterative estimation method that does not require learning from pose-annotated data. In addition, we adopt a novel mesh-based object active shape model that the previous literature has not explored. Our algorithm, ShapeICP, is based on the iterative closest point (ICP) algorithm but is equipped with additional features for the category-level pose and shape estimation task. Although not using pose-annotated data, ShapeICP surpasses many data-driven approaches that rely on pose data for training, opening up a new solution space for researchers to consider. | [
{
"version": "v1",
"created": "Fri, 23 Aug 2024 15:12:55 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 06:56:49 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhang",
"Yihao",
""
],
[
"Sawhney",
"Harpreet S.",
""
],
[
"Leonard",
"John J.",
""
]
] |
2408.13507 | Suthee Ruangwises | Suthee Ruangwises | Tatami Printer: Physical ZKPs for Tatami Puzzles | This paper has appeared at CIAC 2025 | null | 10.1007/978-3-031-92932-8_8 | null | cs.CR cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tatami puzzles are pencil puzzles with an objective to partition a rectangular grid into rectangular regions such that no four regions share a corner point, as well as satisfying other constraints. In this paper, we develop a physical card-based protocol called Tatami printer that can help verify solutions of Tatami puzzles. We then use the Tatami printer to construct zero-knowledge proof protocols for two such puzzles: Tatamibari and Square Jam. These protocols enable a prover to show a verifier the existence of the puzzles' solutions without revealing them. | [
{
"version": "v1",
"created": "Sat, 24 Aug 2024 08:02:52 GMT"
},
{
"version": "v2",
"created": "Sun, 09 Feb 2025 08:42:04 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Ruangwises",
"Suthee",
""
]
] |
2408.14732 | Bojun Xiong | Bojun Xiong, Si-Tong Wei, Xin-Yang Zheng, Yan-Pei Cao, Zhouhui Lian, Peng-Shuai Wang | OctFusion: Octree-based Diffusion Models for 3D Shape Generation | Accepted to Computer Graphics Forum (SGP), 2025 | null | null | null | cs.CV cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models have emerged as a popular method for 3D generation. However, it is still challenging for diffusion models to efficiently generate diverse and high-quality 3D shapes. In this paper, we introduce OctFusion, which can generate 3D shapes with arbitrary resolutions in 2.5 seconds on a single Nvidia 4090 GPU, and the extracted meshes are guaranteed to be continuous and manifold. The key components of OctFusion are the octree-based latent representation and the accompanying diffusion models. The representation combines the benefits of both implicit neural representations and explicit spatial octrees and is learned with an octree-based variational autoencoder. The proposed diffusion model is a unified multi-scale U-Net that enables weights and computation sharing across different octree levels and avoids the complexity of widely used cascaded diffusion schemes. We verify the effectiveness of OctFusion on the ShapeNet and Objaverse datasets and achieve state-of-the-art performances on shape generation tasks. We demonstrate that OctFusion is extendable and flexible by generating high-quality color fields for textured mesh generation and high-quality 3D shapes conditioned on text prompts, sketches, or category labels. Our code and pre-trained models are available at https://github.com/octree-nn/octfusion. | [
{
"version": "v1",
"created": "Tue, 27 Aug 2024 01:55:40 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 16:51:04 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Xiong",
"Bojun",
""
],
[
"Wei",
"Si-Tong",
""
],
[
"Zheng",
"Xin-Yang",
""
],
[
"Cao",
"Yan-Pei",
""
],
[
"Lian",
"Zhouhui",
""
],
[
"Wang",
"Peng-Shuai",
""
]
] |
2408.15323 | Victor Montenegro | Victor Montenegro, Chiranjib Mukhopadhyay, Rozhin Yousefjani, Saubhik Sarkar, Utkarsh Mishra, Matteo G. A. Paris, Abolfazl Bayat | Review: Quantum Metrology and Sensing with Many-Body Systems | Review article. Version close to publication | Physics Reports 1134, 1-62 (2025) | 10.1016/j.physrep.2025.05.005 | null | quant-ph cond-mat.stat-mech cond-mat.str-el physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The main power of quantum sensors is achieved when the probe is composed of several particles. In this situation, quantum features such as entanglement contribute to enhancing the precision of quantum sensors beyond the capacity of classical sensors. Originally, quantum sensing was formulated for non-interacting particles that are prepared in a special form of maximally entangled states. These probes are extremely sensitive to decoherence, and any interaction between particles is detrimental to their performance. An alternative framework for quantum sensing has been developed exploiting quantum many-body systems, where the interaction between particles plays a crucial role. In this review, we investigate different aspects of the latter approach for quantum metrology and sensing. Many-body probes have been used in both equilibrium and non-equilibrium scenarios. Quantum criticality has been identified as a resource for achieving quantum-enhanced sensitivity in both scenarios. In equilibrium, various types of criticalities, such as first-order, second-order, topological, and localization phase transitions, have been exploited for sensing purposes. In non-equilibrium scenarios, quantum-enhanced sensitivity has been discovered for Floquet, dissipative, and time crystal phase transitions. While each type of these criticalities has its own characteristics, the presence of one feature is crucial for achieving quantum-enhanced sensitivity: the energy/quasi-energy gap closing. In non-equilibrium quantum sensing, time is another parameter that can affect the sensitivity of the probe. Typically, the sensitivity enhances as the probe evolves in time. In general, a more complete understanding of resources for non-equilibrium quantum sensors is now rapidly evolving. In this review, we provide an overview of recent progress in quantum metrology and sensing using many-body systems. | [
{
"version": "v1",
"created": "Tue, 27 Aug 2024 18:00:01 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Oct 2024 09:46:05 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 09:44:13 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Montenegro",
"Victor",
""
],
[
"Mukhopadhyay",
"Chiranjib",
""
],
[
"Yousefjani",
"Rozhin",
""
],
[
"Sarkar",
"Saubhik",
""
],
[
"Mishra",
"Utkarsh",
""
],
[
"Paris",
"Matteo G. A.",
""
],
[
"Bayat",
"Abolfazl",
""
]
] |
2408.15377 | Dor Minzer | Amey Bhangale, Subhash Khot and Dor Minzer | On Approximability of Satisfiable k-CSPs: V | null | null | null | null | cs.CC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a framework of algorithm vs. hardness for all Max-CSPs and demonstrate it for a large class of predicates. This framework extends the work of Raghavendra [STOC, 2008], who showed a similar result for almost satisfiable Max-CSPs.
Our framework is based on a new hybrid approximation algorithm, which uses a combination of the Gaussian elimination technique (i.e., solving a system of linear equations over an Abelian group) and the semidefinite programming relaxation. We complement our algorithm with a matching dictator vs. quasirandom test that has perfect completeness.
The analysis of our dictator vs. quasirandom test is based on a novel invariance principle, which we call the mixed invariance principle. Our mixed invariance principle is an extension of the invariance principle of Mossel, O'Donnell and Oleszkiewicz [Annals of Mathematics, 2010] which plays a crucial role in Raghavendra's work. The mixed invariance principle allows one to relate 3-wise correlations over discrete probability spaces with expectations over spaces that are a mixture of Guassian spaces and Abelian groups, and may be of independent interest. | [
{
"version": "v1",
"created": "Tue, 27 Aug 2024 19:30:26 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 11:46:35 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Bhangale",
"Amey",
""
],
[
"Khot",
"Subhash",
""
],
[
"Minzer",
"Dor",
""
]
] |
2408.15889 | Adamu Issifu | Adamu Issifu, Franciele M. da Silva, Luis C. N. Santos, D\'ebora P. Menezes and Tobias Frederico | Strongly Interacting Quark Matter in Massive Quark Stars | 21 pages, 9 figures. Accepted version in Class. Quantum Grav | null | null | null | nucl-th hep-ph | http://creativecommons.org/licenses/by/4.0/ | This paper investigates the properties of strongly coupled matter at high baryon densities (\(\rho_B\)) in quark stars (QSs). The QS model is based on the density-dependent quark mass (DDQM) framework, modified (MDDQM) by enhancing the single-gluon interaction to generate higher repulsive pressure. The model parameters are constrained using Bayesian inference, incorporating observational data from the pulsars HESS J1731$-$347, PSR J0030$+$0451, PSR J0740$+$6620, and PSR J0952$-$0607. Our results show that the MDDQM model produces QSs with higher mass and compactness compared to the DDQM model. Among the four MDDQM parameterizations studied, two yield maximum star masses of 1.86 and 2.10 \(\rm M_\odot\) and exhibit near-conformal behavior in the underlying quark matter (QM). The other two parameterizations, yielding QS masses of 2.30 and 2.37 \(\rm M_\odot\), correspond to a stronger interaction in the underlying QM. These findings provide important insights into the equation of state of deconfined QM and its implications for the structure and stability of QSs. | [
{
"version": "v1",
"created": "Wed, 28 Aug 2024 16:02:00 GMT"
},
{
"version": "v2",
"created": "Thu, 12 Dec 2024 14:51:07 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 11:57:03 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Issifu",
"Adamu",
""
],
[
"da Silva",
"Franciele M.",
""
],
[
"Santos",
"Luis C. N.",
""
],
[
"Menezes",
"Débora P.",
""
],
[
"Frederico",
"Tobias",
""
]
] |
2408.15920 | Jan Szalankiewicz | Jan Szalankiewicz, Cristina Martinez-Torres, Wilhelm Stannat | Nonlinear Filtering and Spatial Asymptotic Consistency for SPDEs Observed via Spatio-Temporal Point Processes | Fixed several typos throughout the manuscript, substantially revised Section 4 with improved theoretical bounds, and updated simulations with corresponding code base improvements | null | null | null | math.ST math.PR stat.TH | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we develop the mathematical framework for filtering problems arising from biophysical applications where data is collected from confocal laser scanning microscopy recordings of the space-time evolution of intracellular wave dynamics of biophysical quantities. In these applications, signals are described by stochastic partial differential equations (SPDEs) and observations can be modelled as functionals of marked point processes whose intensities depend on the underlying signal. We derive both the unnormalized and normalized filtering equations for these systems, demonstrate the asymptotic consistency and approximations of finite dimensional observation schemes respectively partial observations. Our theoretical results are validated through extensive simulations using synthetic and real data. These findings contribute to a deeper understanding of filtering with point process observations and provide a robust framework for future research in this area. | [
{
"version": "v1",
"created": "Wed, 28 Aug 2024 16:34:11 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 14:20:01 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Szalankiewicz",
"Jan",
""
],
[
"Martinez-Torres",
"Cristina",
""
],
[
"Stannat",
"Wilhelm",
""
]
] |
2408.16220 | Yiming Zhu | Yiming Zhu, Wenchao Huang, Yan Xiong | Place Protections at the Right Place: Targeted Hardening for Cryptographic Code against Spectre v1 | Accepted to appear at USENIX Security 2025 | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spectre v1 attacks pose a substantial threat to security-critical software, particularly cryptographic implementations. Existing software mitigations, however, often introduce excessive overhead by indiscriminately hardening instructions without assessing their vulnerability. We propose an analysis framework that employs a novel fixpoint algorithm to detect Spectre vulnerabilities and apply targeted hardening. The fixpoint algorithm accounts for program behavior changes induced by stepwise hardening, enabling precise, sound and efficient vulnerability detection. This framework also provides flexibility for diverse hardening strategies and attacker models, enabling customized targeted hardening. We instantiate the framework as LightSLH, which hardens program with provable security.
We evaluate LightSLH on cryptographic algorithms from OpenSSL, Libsodium, NaCL and PQClean. Across all experimental cases, LightSLH provides the lowest overhead among current provable protection strategies, including 0\% overhead in 50\% cases. Notably, the analysis of LightSLH reveals two previously unknown security issues: (1) The compiler can introduce risks overlooked by LLSCT, a hardening method proven secure at the LLVM IR level. We successfully construct a side channel by exploiting compiler-inserted stack loads, confirming this risk. (2) Memory access patterns generated by the scatter-gather algorithm still depend on secrets, even for observers with cache line granularity. These findings and results highlight the importance of applying accurate protections to specific instructions. | [
{
"version": "v1",
"created": "Thu, 29 Aug 2024 02:31:28 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 09:30:12 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhu",
"Yiming",
""
],
[
"Huang",
"Wenchao",
""
],
[
"Xiong",
"Yan",
""
]
] |
2408.16617 | Wen Zheng | Xiang Deng, Wen Zheng, Xudong Liao, Haoyu Zhou, Yangyang Ge, Jie Zhao, Dong Lan, Xinsheng Tan, Yu Zhang, Shaoxiong Li, and Yang Yu | Long-Range $ZZ$ Interaction via Resonator-Induced Phase in Superconducting Qubits | 7 pages, 4 figures | Phys. Rev. Lett. 134, 020801 (2025) | 10.1103/PhysRevLett.134.020801 | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Superconducting quantum computing emerges as one of leading candidates for achieving quantum advantage. However, a prevailing challenge is the coding overhead due to limited quantum connectivity, constrained by nearest-neighbor coupling among superconducting qubits. Here, we propose a novel multimode coupling scheme using three resonators driven by two microwaves, based on the resonator-induced phase gate, to extend the $ZZ$ interaction distance between qubits. We demonstrate a CZ gate fidelity exceeding 99.9\% within 160 ns at free spectral range (FSR) of 1.4 GHz, and by optimizing driving pulses, we further reduce the residual photon to nearly $10^{-3}$ within 100 ns at FSR of 0.2 GHz. These facilitate the long-range CZ gate over separations reaching sub-meters, thus significantly enhancing qubit connectivity and making a practical step towards the scalable integration and modularization of quantum processors. Specifically, our approach supports the implementation of quantum error correction codes requiring high connectivity, such as low-density parity check codes that paves the way to achieving fault-tolerant quantum computing. | [
{
"version": "v1",
"created": "Thu, 29 Aug 2024 15:24:08 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Sep 2024 10:32:11 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Deng",
"Xiang",
""
],
[
"Zheng",
"Wen",
""
],
[
"Liao",
"Xudong",
""
],
[
"Zhou",
"Haoyu",
""
],
[
"Ge",
"Yangyang",
""
],
[
"Zhao",
"Jie",
""
],
[
"Lan",
"Dong",
""
],
[
"Tan",
"Xinsheng",
""
],
[
"Zhang",
"Yu",
""
],
[
"Li",
"Shaoxiong",
""
],
[
"Yu",
"Yang",
""
]
] |
2409.00532 | Michael K. -H. Kiessling | Michael K.-H. Kiessling, Boris L. Altshuler, and Emil A. Yuzbashyan | Bounds on $T_c$ in the Eliashberg theory of Superconductivity. II: Dispersive phonons | 44 pages, 2 figures, revised and corrected version (in particular, the proof of Prop.4 has been corrected); accepted for publication in J. Statist. Phys | null | null | null | math-ph cond-mat.supr-con math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The standard Eliashberg theory of superconductivity is studied, in which the effective electron-electron interactions are mediated by generally dispersive phonons, with Eliashberg spectral function $\alpha^2 F(\omega)\geq 0$ that is $\propto\omega^2$ for small $\omega>0$ and vanishes for large $\omega$. The Eliashberg function also defines the electron-phonon coupling strength $\lambda:= 2 \int_0^\infty\frac{\alpha^2 F(\omega)}{\omega}d\omega$. Setting $\frac{2\alpha^2 F(\omega)}{\omega}d\omega =: \lambda P(d\omega)$, formally defining a probability measure $P(d\omega)$ with compact support, and assuming as usual that the phase transition between normal and superconductivity coincides with the linear stability boundary $\mathscr{S}_{\!c}$ of the normal region against perturbations toward the superconducting region, it is shown that $\mathscr{S}_{\!c}$ is a graph of a function $\Lambda(P,T)$ that is determined by a variational principle: if $(\lambda,P,T)\in\mathscr{S}_{\!c}$, then $\lambda = 1/\mathfrak{k}(P,T)$, where $\mathfrak{k}(P,T)>0$ is the largest eigenvalue of a compact self-adjoint operator $\mathfrak{K}(P,T)$ on $\ell^2$ sequences constructed in the paper. Given $P$, sufficient conditions on $T$ are stated under which the map $T\mapsto \lambda = \Lambda(P,T)$ is invertible. For sufficiently large $\lambda$ this yields: (i) the existence of a critical temperature $T_c$ as function of $\lambda$ and $P$; (ii) a sequence of lower bounds on $T_c(\lambda,P)$ that converges to $T_c(\lambda,P)$. Also obtained is an upper bound on $T_c(\lambda,P)$. It agrees with the asymptotic form $T_c(\lambda,P) \sim C \sqrt{\langle \omega^2\rangle} \sqrt{\lambda}$ valid for $\lambda\sim\infty$, given $P$, though with a constant $C$ that is a factor $\approx 2.034$ larger than the sharp constant. Here, $\langle\omega^2\rangle := \int_0^\infty \omega^2 P(d\omega)$. | [
{
"version": "v1",
"created": "Sat, 31 Aug 2024 19:43:03 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 13:01:15 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kiessling",
"Michael K. -H.",
""
],
[
"Altshuler",
"Boris L.",
""
],
[
"Yuzbashyan",
"Emil A.",
""
]
] |
2409.00533 | Michael K. -H. Kiessling | Michael K.-H. Kiessling, Boris L. Altshuler, and Emil A. Yuzbashyan | Bounds on $T_c$ in the Eliashberg theory of Superconductivity. I: The $\gamma$-model | 49 pages, 2 figures, revised preprint version; appeared with different layout in J. Statist. Phys | Journal of Statistical Physics, vol. 192, art.69, 35pp. (2025) | 10.1007/s10955-025-03446-5 | null | math-ph cond-mat.supr-con math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using the recent reformulation for the Eliashberg theory of superconductivity in terms of a classical interacting Bloch spin chain model, rigorous upper and lower bounds on the critical temperature $T_c$ are obtained for the $\gamma$ model -- a version of Eliashberg theory in which the effective electron-electron interaction is proportional to $(g/|\omega_n-\omega_m|)^{\gamma}$, where $\omega_n-\omega_m$ is the transferred Matsubara frequency, $g>0$ a reference energy, and $\gamma>0$ a parameter. The rigorous lower bounds are based on a variational principle that identifies $(T_c/g)^\gamma$ with the largest (positive) eigenvalue of an explicitly constructed compact, self-adjoint operator $\mathfrak{G}(\gamma)$. These lower bounds form an increasing sequence that converges to $T_c(g,\gamma)$. The upper bound on $T_c(g,\gamma)$ is based on fixed point theory, proving linear stability of the normal state for $T$ larger than the upper bound on $T_c(g,\gamma)$. | [
{
"version": "v1",
"created": "Sat, 31 Aug 2024 19:43:59 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 12:35:38 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kiessling",
"Michael K. -H.",
""
],
[
"Altshuler",
"Boris L.",
""
],
[
"Yuzbashyan",
"Emil A.",
""
]
] |
2409.01667 | Muye Huang | Muye Huang, Lingling Zhang, Lai Han, Wenjun Wu, Xinyu Zhang, Jun Liu | VProChart: Answering Chart Question through Visual Perception Alignment Agent and Programmatic Solution Reasoning | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Charts are widely used for data visualization across various fields, including education, research, and business. Chart Question Answering (CQA) is an emerging task focused on the automatic interpretation and reasoning of data presented in charts. However, chart images are inherently difficult to interpret, and chart-related questions often involve complex logical and numerical reasoning, which hinders the performance of existing models. This paper introduces VProChart, a novel framework designed to address these challenges in CQA by integrating a lightweight Visual Perception Alignment Agent (VPAgent) and a Programmatic Solution Reasoning approach. VPAgent aligns and models chart elements based on principles of human visual perception, enhancing the understanding of chart context. The Programmatic Solution Reasoning approach leverages large language models (LLMs) to transform natural language reasoning questions into structured solution programs, facilitating precise numerical and logical reasoning. Extensive experiments on benchmark datasets such as ChartQA and PlotQA demonstrate that VProChart significantly outperforms existing methods, highlighting its capability in understanding and reasoning with charts. | [
{
"version": "v1",
"created": "Tue, 03 Sep 2024 07:19:49 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 07:55:50 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Huang",
"Muye",
""
],
[
"Zhang",
"Lingling",
""
],
[
"Han",
"Lai",
""
],
[
"Wu",
"Wenjun",
""
],
[
"Zhang",
"Xinyu",
""
],
[
"Liu",
"Jun",
""
]
] |
2409.02121 | Michael K. -H. Kiessling | Michael K.-H. Kiessling, Boris L. Altshuler, and Emil A. Yuzbashyan | Bounds on $T_c$ in the Eliashberg theory of Superconductivity. III: Einstein phonons | 36 pages, 2 figures, revised version, accepted for publication in J. Statist. Phys | null | null | null | cond-mat.supr-con math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The dispersionless limit of the standard Eliashberg theory of superconductivity is studied. The effective electron-electron interactions are mediated by Einstein phonons of frequency $\Omega>0$, equipped with electron-phonon coupling strength $\lambda$. This allows for a detailed evaluation of the general results on $T_c$ for phonons with non-trivial dispersion relation, obtained in a previous paper, (II), by the authors. The variational principle for the linear stability boundary $\mathscr{S}_{\!c}$ of the normal state region against perturbations toward the superconducting region, obtained in (II), simplifies as follows: If $(\lambda,\Omega,T)\in\mathscr{S}_{\!c}$, then $\lambda = 1/\mathfrak{h}(\varpi)$, where $\varpi:=\Omega/2\pi T$, and where $\mathfrak{h}(\varpi)>0$ is the largest eigenvalue of a compact self-adjoint operator $\mathfrak{H}(\varpi)$ on $\ell^2$ sequences; $\mathfrak{H}(\varpi)$ is the dispersionless limit $P(d\omega)\to\delta(\omega-\Omega)d\omega$ of the operator $\mathfrak{K}(P,T)$ of (II). It is shown that when $\varpi \leq \sqrt{2}$, then the map $\varpi\mapsto\mathfrak{h}(\varpi)$ is invertible. For $\lambda>0.77$ this yields: (i) the existence of a critical temperature $T_c(\lambda,\Omega) = \Omega f(\lambda)$; (ii) a sequence of lower bounds on $f(\lambda)$ that converges to $f(\lambda)$. Also obtained is an upper bound on $T_c(\lambda,\Omega)$, which agrees with the asymptotic behavior $T_c(\lambda,\Omega) \sim C \Omega\sqrt{\lambda}$ for $\lambda\sim\infty$, given $\Omega$, though with $C\approx 2.034 C_\infty$, where $C_\infty := \frac{1}{2\pi}\mathfrak{k}(2)^\frac12 =0.1827262477...$ is the optimal constant, and $\mathfrak{k}(\gamma)>0$ the largest eigenvalue of a compact self-adjoint operator for the $\gamma$ model, determined in the first paper, (I), on $T_c$ by the authors. | [
{
"version": "v1",
"created": "Sat, 31 Aug 2024 20:05:35 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 13:10:41 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kiessling",
"Michael K. -H.",
""
],
[
"Altshuler",
"Boris L.",
""
],
[
"Yuzbashyan",
"Emil A.",
""
]
] |
2409.02171 | Kai Klocke | Kai Klocke, Daniel Simm, Guo-Yi Zhu, Simon Trebst, Michael Buchhold | Entanglement dynamics in monitored Kitaev circuits: loop models, symmetry classification, and quantum Lifshitz scaling | null | Phys. Rev. B 111, 224301 (2025) | 10.1103/PhysRevB.111.224301 | null | quant-ph cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum circuits offer a versatile platform for simulating digital quantum dynamics and uncovering novel states of non-equilibrium quantum matter. One principal example are measurement-induced phase transitions arising from non-unitary dynamics in monitored circuits, which employ mid-circuit measurements as an essential building block next to standard unitary gates. Although a comprehensive understanding of dynamics in generic circuits is still evolving, we contend that monitored quantum circuits yield robust phases of dynamic matter, which -- akin to Hamiltonian ground state phases -- can be categorized based on symmetries and spatial dimensionality. To illustrate this concept, we focus on quantum circuits within symmetry classes BDI and D, which are measurement-only adaptations of the paradigmatic Kitaev and Yao-Kivelson models, embodying particle-hole-symmetric Majorana fermions with or without time-reversal. We establish a general framework -- Majorana loop models -- for both symmetry classes to provide access to the phenomenology of the entanglement dynamics in these circuits, displaying both an area-law phase of localized Majorana loops and a delocalized, highly entangled Majorana liquid phase. The two phases are separated by a continuous transition displaying quantum Lifshitz scaling, albeit with critical exponents of two distinct universality classes. The loop model framework provides not only analytical understanding of these universality classes in terms of non-linear sigma models, but also allows for highly efficient numerical techniques capable of simulating excessively large circuits with up to $10^8$ qubits. We utilize this framework to accurately determine universal probes that distinguish both the entangled phases and the critical points of the two symmetry classes. Our work thereby further solidifies the concept of emergent circuit phases and their phase transitions. | [
{
"version": "v1",
"created": "Tue, 03 Sep 2024 18:00:01 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Klocke",
"Kai",
""
],
[
"Simm",
"Daniel",
""
],
[
"Zhu",
"Guo-Yi",
""
],
[
"Trebst",
"Simon",
""
],
[
"Buchhold",
"Michael",
""
]
] |
2409.02516 | Chao Liu | Chao Liu | The emergence of nonlinear Jeans-type instabilities for quasilinear wave equations | 54 pages, multiple places improved and typos fixed | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article contributes a key ingredient to the longstanding open problem of understanding the fully nonlinear version of Jeans instability, as highlighted by A. Rendall [Living Rev. Relativ. 8, 6 (2005)]. We establish a family of self-increasing blowup solutions for the following class of quasilinear wave equations (a model of the Peebles' and Noh-Hwang's equations) that have not previously been studied: \[
\partial^2_t \varrho- \biggl(\frac{ \mathsf{m}^2 (\partial_{t}\varrho )^2}{(1+\varrho )^2} + 4(\mathsf{k}-\mathsf{m}^2)(1+\varrho )\biggr) \Delta \varrho = F(t,\varrho,\partial_{\mu} \varrho) \] where $F$ is given by \[
F(t,\varrho,\partial_{\mu} \varrho):=
\underbrace{\frac{2}{3 } \varrho (1+\varrho) }_{ \text{(i) self-increasing}} \underbrace{-\frac{1}{3} \partial_{t}\varrho }_{ \text{(ii) damping}}
+ \underbrace{\frac{4}{3} \frac{(\partial_{t}\varrho )^2}{1+\varrho } }_{\text{(iii) Riccati}} + \underbrace{ \biggl(\mathsf{m}^2 \frac{ (\partial_{t}\varrho )^2}{(1+\varrho )^2} + 4(\mathsf{k}-\mathsf{m}^2) (1+\varrho ) \biggr) q^i \partial_{i}\varrho }_{\text{(iv) convection}} - \mathtt{K}^{ij} \partial_{i}\varrho\partial_{j}\varrho. \] The result implies the solutions can attain arbitrarily large values over time, leading to self-increasing singularities at some future endpoints of null geodesics provided the inhomogeneous perturbations of data are sufficiently small. Moreover, the solution exhibits almost blowup behavior in the long-wavelength domain. This phenomenon is referred to as the nonlinear Jeans-type instability because this wave equation is closely related to the nonlinear version of the Jeans instability problem in the Euler-Poisson and Einstein-Euler systems, which characterizes the formation of nonlinear structures in the universe. The growth rate of $\varrho$ is significantly faster than that of the solutions to the classical linearized Jeans instability. | [
{
"version": "v1",
"created": "Wed, 04 Sep 2024 08:25:51 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Jun 2025 00:54:33 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Liu",
"Chao",
""
]
] |
2409.02659 | Nuno Crokidakis | Nuno Crokidakis | Dynamics of drug trafficking: Results from a simple compartmental model | 17 pages, 6 figures, to appear in IJMPC | Int. J. Mod. Phys. C 36, 2450201 (2025) | 10.1142/S0129183124502012 | null | physics.soc-ph cs.SI | http://creativecommons.org/publicdomain/zero/1.0/ | In this work we propose a simple model for the emergence of drug dealers. For this purpose, we built a compartmental model considering four subpopulations, namely susceptibles, passive supporters, drug dealers and arrested drug dealers. The target is to study the influence of the passive supporters on the long-time prevalence of drug dealers. Passive supporters are people who are passively consenting to the drug trafficking cause. First we consider the model on a fully-connected newtork, in such a way that we can write a rate equation for each subpopulation. Our analytical and numerical results show that the emergence of drug dealers is a consequence of the rapid increase number of passive supporters. Such increase is associated with a nonequilibrium active-absorbing phase transition. After that, we consider the model on a two-dimensional square lattice, in order to compare the results in the presence of a simple social network with the previous results. The Monte Carlo simulation results suggest a similar behavior in comparison with the fully-connected network case, but the location of the critical point of the transition is distinct, due to the neighbors' correlations introduced by the presence of the lattice. | [
{
"version": "v1",
"created": "Wed, 04 Sep 2024 12:33:55 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Crokidakis",
"Nuno",
""
]
] |
2409.02799 | Mark V. Kartsovnik | Shamil Erkenov, Sergej Fust, Sebastian Oberbauer, Werner Biberacher, Natalia D. Kushch, Harald Mueller, Francis L. Pratt, Rudolf Gross, and Mark V. Kartsovnik | Electronic correlations and spin frustration in the molecular conductors $\kappa$-(BEDT-TTF)$_2$X probed by magnetic quantum oscillations | 13 pages,including 7 Figures. Supplemental Material is available upon request | Phys. Rev B 110, 205149 (2024) | 10.1103/PhysRevB.110.205149 | null | cond-mat.str-el | http://creativecommons.org/publicdomain/zero/1.0/ | The layered molecular conductors $\kappa$-(BEDT-TTF)$_2$X are a perfect experimental platform for studying the physics of the Mott transition and related exotic electronic states. In these materials, the subtle balance between various instabilities of the normal metallic state can be efficiently changed by applying a very moderate external pressure or by subtle chemical modifications, e.g. by a replacement of the insulating anion X$^{-}$, frequently referred to as ``chemical pressure''. A crucially important but still unsettled issue is an exact understanding of the influence of physical and chemical pressure on the electronic structure. Here, we use magnetic quantum oscillations to explore in a broad pressure range the behavior of the key parameters governing the Mott physics, the electronic correlation strength ratio $U/t$ and the spin frustration ratio $t'/t$ in two $\kappa$ salts, the ambient-pressure antiferromagnetic insulator with X = Cu[N(CN)$_2$]Cl and the ambient-pressure superconductor with X = Cu(NCS)$_2$. Our analysis shows that pressure effectively changes not only the conduction bandwidth but also the degree of spin frustration, thus weakening both the electronic correlation strength and the magnetic ordering instability. At the same time, we find that the replacement of the anion Cu[N(CN)$_2$]Cl$^-$ by Cu(NCS)$_2^-$ results in a significant increase of the frustration parameter $t'/t$, leaving the correlation strength essentially unchanged. | [
{
"version": "v1",
"created": "Wed, 04 Sep 2024 15:16:48 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Erkenov",
"Shamil",
""
],
[
"Fust",
"Sergej",
""
],
[
"Oberbauer",
"Sebastian",
""
],
[
"Biberacher",
"Werner",
""
],
[
"Kushch",
"Natalia D.",
""
],
[
"Mueller",
"Harald",
""
],
[
"Pratt",
"Francis L.",
""
],
[
"Gross",
"Rudolf",
""
],
[
"Kartsovnik",
"Mark V.",
""
]
] |
2409.02848 | Hongye Yu | Hongye Yu, Tzu-Chieh Wei | The robustness condition for general disordered discrete time crystals, and subspace-thermal DTCs from phase transitions between different n-tuple DTCs | 32 pages, 11 figures, update DOI | Physical Review B 111.17 (2025): 174311 | 10.1103/PhysRevB.111.174311 | null | quant-ph cond-mat.dis-nn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a new Floquet time crystal model that responds in arbitrary multiples of the driving period. Such an $n$-tuple discrete time crystal is theoretically constructed by permuting spins in a disordered chain and is well suited for experimental implementations. Transitions between these time crystals with different periods give rise to a novel phase of matter that we call subspace-thermal discrete time crystals, where states within subspaces of definite charges are fully thermalized at an early time. However, the whole system still robustly responds to the periodic driving subharmonically, with a period being the greatest common divisor of the original two periods. Existing theoretical analysis from many-body localization cannot be used to understand the rigidity of such subspace-thermal time crystal phases. To resolve this, we develop a new theoretical framework for the robustness of DTCs from the perspective of the robust $2\pi/n$ quasi-energy gap. Its robustness is rigorously proved if the system satisfies a certain condition where the mixing length, defined by the Hamming distance of the symmetry charges, does not exceed a global threshold. The proof applies beyond the models considered here to other existing DTCs realized by kicking disordered systems, where conventional MBL-DTCs can be regarded as a special case of the subspace-thermal DTC with the subspace dimension being one, thus offering a systematic way to construct new discrete time crystal models. We also introduce the notion of DTC-charges that allow us to probe the observables that spontaneously break the time-translation symmetry in both the regular discrete time crystals and subspace-thermal discrete time crystals. Moreover, our discrete time crystal models can be generalized to systems with higher spin magnitudes or qudits, as well as to higher spatial dimensions. | [
{
"version": "v1",
"created": "Wed, 04 Sep 2024 16:19:43 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Sep 2024 20:17:44 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Oct 2024 22:49:40 GMT"
},
{
"version": "v4",
"created": "Fri, 07 Mar 2025 20:58:36 GMT"
},
{
"version": "v5",
"created": "Tue, 29 Apr 2025 17:13:33 GMT"
},
{
"version": "v6",
"created": "Sat, 07 Jun 2025 01:42:54 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yu",
"Hongye",
""
],
[
"Wei",
"Tzu-Chieh",
""
]
] |
2409.02941 | Xue-Ping Wang | Chen Meng, Yun-Mao Zhang, Xue-ping Wang | The characterizations of monotone functions which generate associative functions | 16 | Fuzzy Sets and Systems 500 (2025) 109201 | 10.1016/j.fss.2024.109201 | 15 | math.GM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Associativity of a two-place function $T: [0,1]^2\rightarrow [0,1]$ defined by $T(x,y)=f^{(-1)}(F(f(x),f(y)))$ where $F:[0,\infty]^2\rightarrow[0,\infty]$ is an associative function, $f: [0,1]\rightarrow [0,\infty]$ is a monotone function which satisfies either $f(x)=f(x^{+})$ when $f(x^{+})\in \mbox{Ran}(f)$ or $f(x)\neq f(y)$ for any $y\neq x$ when $f(x^{+})\notin \mbox{Ran}(f)$ for all $x\in[0,1]$ and $f^{(-1)}:[0,\infty]\rightarrow[0,1]$ is a pseudo-inverse of $f$ depends only on properties of the range of $f$. The necessary and sufficient conditions for the $T$ to be associative are presented by applying the properties of the monotone function $f$. | [
{
"version": "v1",
"created": "Sat, 24 Aug 2024 01:56:19 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Meng",
"Chen",
""
],
[
"Zhang",
"Yun-Mao",
""
],
[
"Wang",
"Xue-ping",
""
]
] |
2409.03567 | Bruno Degli Esposti | Oleg Davydov and Bruno Degli Esposti | Meshless moment-free quadrature formulas arising from numerical differentiation | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We suggest a method for simultaneously generating high order quadrature weights for integrals over Lipschitz domains and their boundaries that requires neither meshing nor moment computation. The weights are determined on pre-defined scattered nodes as a minimum norm solution of a sparse underdetermined linear system arising from a discretization of a suitable boundary value problem by either collocation or meshless finite differences. The method is easy to implement independently of the domain's representation, since it only requires as inputs the position of all quadrature nodes and the direction of outward-pointing normals at each node belonging to the boundary. Numerical experiments demonstrate the robustness and high accuracy of the method on a number of smooth and piecewise smooth domains in 2D and 3D, including some with reentrant corners and edges. Comparison with quadrature schemes provided by the state-of-the-art open source packages Gmsh and MFEM shows that the new method is competitive in terms of accuracy for a given number of nodes. | [
{
"version": "v1",
"created": "Thu, 05 Sep 2024 14:21:26 GMT"
},
{
"version": "v2",
"created": "Mon, 06 Jan 2025 10:49:12 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 02:58:17 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Davydov",
"Oleg",
""
],
[
"Esposti",
"Bruno Degli",
""
]
] |
2409.03786 | Andrzej Krasi\'nski | Andrzej Krasi\'nski | Causality in the maximally extended Reissner--Nordstr\"{o}m spacetime with identifications | 34 pages, 13 figures. Two small typos on pages 22 and 24 of the previous version corrected. Matches the published version | Reports on Mathematical Physics 95, 181 - 214 (2025) | null | null | physics.gen-ph gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The maximally extended Reissner--Nordstr\"{o}m (RN) spacetime with $e^2 < m^2$ can be interpreted either as an infinite chain of asymptotically flat regions connected by tunnels between timelike singularities or as a set of just one pair of asymptotically flat regions and one tunnel; the repetitions of this set in the infinite chain being identified. The second interpretation gives rise to the suspicion of acausality, i.e. the possibility of sending messages to one's own past. A numerical investigation of this problem was carried out in this paper and gave the following result. Let E be the initial point of a radial timelike future-directed ingoing geodesic G, lying halfway between the outer horizon and the image of the null infinity in the maximally extended RN spacetime. Let E$'$ be the first future copy of E. It was verified whether the turning point of G lies outside or inside the past light cone (PLC) of E$'$. In the second case the breach of causality does occur. It turned out that the acausality is present when $V_E$, the timelike coordinate of E, is negative with a sufficiently large $|V_E|$, and is absent with a sufficiently large $V_E > 0$. In between these values there exists a $\widetilde{V}_E$, dependent on the initial data for the geodesic, for which the turning point lies on the PLC. So, the identification does lead to acausality. Nonradial timelike and null geodesics were also investigated, and a few hitherto unknown properties of the maximal extension were revealed. For example, the singularity arc at $r = 0$ may be convex or concave, depending on the values of $m$ and $e$. | [
{
"version": "v1",
"created": "Fri, 30 Aug 2024 15:47:42 GMT"
},
{
"version": "v2",
"created": "Tue, 08 Oct 2024 13:09:31 GMT"
},
{
"version": "v3",
"created": "Mon, 10 Mar 2025 18:24:23 GMT"
},
{
"version": "v4",
"created": "Thu, 03 Apr 2025 13:18:11 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Krasiński",
"Andrzej",
""
]
] |
2409.03788 | Cheng Qian | Cheng Qian and Hainan Zhang and Lei Sha and Zhiming Zheng | HSF: Defending against Jailbreak Attacks with Hidden State Filtering | WWW2025 WSAI BESTPAPER | null | null | null | cs.CR cs.AI cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the growing deployment of LLMs in daily applications like chatbots and content generation, efforts to ensure outputs align with human values and avoid harmful content have intensified. However, increasingly sophisticated jailbreak attacks threaten this alignment, aiming to induce unsafe outputs. Current defense efforts either focus on prompt rewriting or detection, which are limited in effectiveness due to the various design of jailbreak prompts, or on output control and detection, which are computationally expensive as they require LLM inference. Therefore, designing a pre-inference defense method that resists diverse jailbreak prompts is crucial for preventing LLM jailbreak attacks. We observe that jailbreak attacks, safe queries, and harmful queries exhibit different clustering patterns within the LLM's hidden state representation space. This suggests that by leveraging the LLM's hidden state representational capabilities, we can analyze the LLM's forthcoming behavior and proactively intervene for defense. In this paper, we propose a jailbreak attack defense strategy based on a Hidden State Filter (HSF), a lossless architectural defense mechanism that enables the model to preemptively identify and reject adversarial inputs before the inference process begins. We activate its defensive potential through an additional plugin module, effectively framing the defense task as a classification problem. Experimental results on two benchmark datasets, utilizing three different LLMs, show that HSF significantly enhances resilience against six cutting-edge jailbreak attacks. It significantly reduces the success rate of jailbreak attacks while minimally impacting responses to benign user queries, with negligible inference overhead, and outperforming defense baselines.Our code and data are available at https://anonymous.4open.science/r/Hidden-State-Filtering-8652/ | [
{
"version": "v1",
"created": "Sat, 31 Aug 2024 06:50:07 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 09:23:17 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Qian",
"Cheng",
""
],
[
"Zhang",
"Hainan",
""
],
[
"Sha",
"Lei",
""
],
[
"Zheng",
"Zhiming",
""
]
] |
2409.04683 | Nour Shaheen | Nour Shaheen, Tamer Elsharnouby, Marwan Torki | C2F-CHART: A Curriculum Learning Approach to Chart Classification | This paper has been accepted for publication in the proceedings of the 2024 International Conference on Pattern Recognition (ICPR) | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | In scientific research, charts are usually the primary method for visually representing data. However, the accessibility of charts remains a significant concern. In an effort to improve chart understanding pipelines, we focus on optimizing the chart classification component. We leverage curriculum learning, which is inspired by the human learning process. In this paper, we introduce a novel training approach for chart classification that utilizes coarse-to-fine curriculum learning. Our approach, which we name C2F-CHART (for coarse-to-fine) exploits inter-class similarities to create learning tasks of varying difficulty levels. We benchmark our method on the ICPR 2022 CHART-Infographics UB UNITEC PMC dataset, outperforming the state-of-the-art results. | [
{
"version": "v1",
"created": "Sat, 07 Sep 2024 02:35:16 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 20:49:24 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Shaheen",
"Nour",
""
],
[
"Elsharnouby",
"Tamer",
""
],
[
"Torki",
"Marwan",
""
]
] |
2409.05101 | Jadson Santos | Jadson Santos and Daniel Alencar da Costa and Shane McIntosh and Uir\'a Kulesza | On the Need to Monitor Continuous Integration Practices -- An Empirical Study | Submitted to the Empirical Software Engineering Journal | null | null | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continuous Integration (CI) encompasses a set of widely adopted practices that enhance software development. However, there are indications that developers may not adequately monitor CI practices. Hence, this paper explores developers' perceptions regarding the monitoring CI practices. To achieve this, we first perform a Document Analysis to assess developers' expressed need for practice monitoring in pull requests comments generated by developers during the development process. After that, we conduct a survey among developers from 121 open-source projects to understand perception of the significance of monitoring seven CI practices in their projects. Finally, we triangulate the emergent themes from our survey by performing a second Document Analysis to understand the extent of monitoring features supported by existing CI services. Our key findings indicate that: 1) the most frequently mentioned CI practice during the development process is ``Test Coverage'' (> 80\%), while ``Build Health'' and ``Time to Fix a Broken Build'' present notable opportunities for monitoring CI practices; 2) developers do not adequately monitor all CI practices and express interest in monitoring additional practices; and 3) the most popular CI services currently offer limited native support for monitoring CI practices, requiring the use of third-party tools. Our results lead us to conclude that monitoring CI practices is often overlooked by both CI services and developers. Using third-party tools in conjunction with CI services is challenging, they monitor some redundant practices and still falls short of fully supporting CI practices monitoring. Therefore, CI services should implement CI practices monitoring, which would facilitate and encourage developers to monitor them. | [
{
"version": "v1",
"created": "Sun, 08 Sep 2024 14:08:07 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 13:15:42 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Santos",
"Jadson",
""
],
[
"da Costa",
"Daniel Alencar",
""
],
[
"McIntosh",
"Shane",
""
],
[
"Kulesza",
"Uirá",
""
]
] |
2409.05129 | Byung-Geel Yu | Tae Keun Choi, Kook-Jin Kong, and Byung-Geel Yu | Nuclear transparency and shadowing of the two-step process in the $A(e,e'\pi^+)$ reaction | 7 pages, 4 figures, 1 table | null | null | null | nucl-th hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate nuclear transparency induced by pion electroproduction on nuclei, $A(e,e'\pi^+)$, with our interest in the role of the two-step process in the Glauber multiple scattering theory. Based on the framework in which the quantum diffusion model is incorporated into the Glauber theory, the experimental data in the range of photon virtuality $0.5< Q^2< 10$ GeV$^2/c^2$ are analyzed, including the recent JLab data at the 6 GeV electron beam together with the proposal for the 12 GeV upgrade. Application of the quantum diffusion model during the formation length of the quark/antiquark $(q\bar{q})$ structure of a pion reproduces the increase in the nuclear transparency up to high $Q^2$, showing evidence of color transparency. In contrast, through the coherent length of the $\gamma^*$ fluctuation into $\rho^0$, the two-step process with the $\rho N$ scattering cross section chosen to be $\sigma_{\rho N}=\sigma_{\pi N}$ contributes to a reduction of the nuclear transparency in the whole range of $Q^2$. Combined with the one-step process (direct photon coupling) usually adopted in the original formulation of the Glauber theory, this reaction mechanism provides another source of the $Q^2$ dependence of the transparency at low virtualities through the shadowing in the initial state interaction. By the further absorption due to the shadowing, a better agreement is obtained with the pion transparency measured in the $A(e,e'\pi^+)$ reaction for $^{12}$C, $^{27}$Al, $^{63}$Cu, and $^{197}$Au nuclei. A discussion of the $Q^2$ dependence of the parameter $\alpha$ for the transparency $T_A=A^{\alpha-1}$ is given. | [
{
"version": "v1",
"created": "Sun, 08 Sep 2024 15:18:42 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Sep 2024 14:31:45 GMT"
},
{
"version": "v3",
"created": "Thu, 27 Mar 2025 17:24:41 GMT"
},
{
"version": "v4",
"created": "Sun, 30 Mar 2025 15:01:04 GMT"
},
{
"version": "v5",
"created": "Sun, 06 Apr 2025 13:56:19 GMT"
},
{
"version": "v6",
"created": "Sun, 13 Apr 2025 15:10:11 GMT"
},
{
"version": "v7",
"created": "Mon, 09 Jun 2025 16:52:28 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Choi",
"Tae Keun",
""
],
[
"Kong",
"Kook-Jin",
""
],
[
"Yu",
"Byung-Geel",
""
]
] |
2409.05191 | Zhiyang Wang | Zhiyang Wang and Juan Cervino and Alejandro Ribeiro | Generalization of Geometric Graph Neural Networks with Lipschitz Loss Functions | 13 pages, 6 figures | null | null | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study the generalization capabilities of geometric graph neural networks (GNNs). We consider GNNs over a geometric graph constructed from a finite set of randomly sampled points over an embedded manifold with topological information captured. We prove a generalization gap between the optimal empirical risk and the optimal statistical risk of this GNN, which decreases with the number of sampled points from the manifold and increases with the dimension of the underlying manifold. This generalization gap ensures that the GNN trained on a graph on a set of sampled points can be utilized to process other unseen graphs constructed from the same underlying manifold. The most important observation is that the generalization capability can be realized with one large graph instead of being limited to the size of the graph as in previous results. The generalization gap is derived based on the non-asymptotic convergence result of a GNN on the sampled graph to the underlying manifold neural networks (MNNs). We verify this theoretical result with experiments on multiple real-world datasets. | [
{
"version": "v1",
"created": "Sun, 08 Sep 2024 18:55:57 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 23:13:03 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Wang",
"Zhiyang",
""
],
[
"Cervino",
"Juan",
""
],
[
"Ribeiro",
"Alejandro",
""
]
] |
2409.05316 | Masahiro Yukawa | Masahiro Yukawa | Continuous Relaxation of Discontinuous Shrinkage Operator: Proximal Inclusion and Conversion | null | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | We present a principled way of deriving a continuous relaxation of a given discontinuous shrinkage operator, which is based on two fundamental results, proximal inclusion and conversion. Using our results, the discontinuous operator is converted, via double inversion, to a continuous operator; more precisely, the associated ``set-valued'' operator is converted to a ``single-valued'' Lipschitz continuous operator. The first illustrative example is the firm shrinkage operator which can be derived as a continuous relaxation of the hard shrinkage operator. We also derive a new operator as a continuous relaxation of the discontinuous shrinkage operator associated with the so-called reversely ordered weighted L1 (ROWL) penalty. Numerical examples demonstrate potential advantages of the continuous relaxation. | [
{
"version": "v1",
"created": "Mon, 09 Sep 2024 04:08:49 GMT"
},
{
"version": "v2",
"created": "Sat, 01 Mar 2025 09:25:44 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 11:31:44 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yukawa",
"Masahiro",
""
]
] |
2409.06056 | Aaron Trigg | Aaron C. Trigg, Rachael Stewart, Alex Van Kooten, Eric Burns, Matthew G. Baring, Dmitry D. Frederiks, Daniela Huppenkothen, Brendan O'Connor, Oliver J. Roberts, Zorawar Wadiasingh, George Younes, Narayana Bhat, Michael S. Briggs, Malte Busmann, Adam Goldstein, Daniel Gruen, Lei Hu, Chryssa Kouveliotou, Michela Negro, Antonella Palmese, Arno Riffeser, Lorenzo Scotton, Dmitry S. Svinkin, Peter Veres, Raphael Z\"oller | Extragalactic Magnetar Giant Flare GRB 231115A: Insights from Fermi/GBM Observations | null | null | 10.1051/0004-6361/202452268 | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | We present the detection and analysis of GRB 231115A, a candidate extragalactic magnetar giant flare (MGF) observed by Fermi/GBM and localized by INTEGRAL to the starburst galaxy M82. This burst exhibits distinctive temporal and spectral characteristics that align with known MGFs, including a short duration and a high peak energy. Gamma-ray analyses reveal significant insights into this burst, supporting conclusions already established in the literature: our time-resolved spectral studies provide further evidence that GRB 231115A is indeed a MGF. Significance calculations also suggest a robust association with M82, further supported by a high Bayes factor that minimizes the probability of chance alignment with a neutron star merger. Despite extensive follow-up efforts, no contemporaneous gravitational wave or radio emissions were detected. The lack of radio emission sets stringent upper limits on possible radio luminosity. Constraints from our analysis show no fast radio bursts (FRBs) associated with two MGFs. X-ray observations conducted post-burst by Swift/XRT and XMM/Newton provided additional data, though no persistent counterparts were identified. Our study underscores the importance of coordinated multi-wavelength follow-up and highlights the potential of MGFs to enhance our understanding of short GRBs and magnetar activities in the cosmos. Current MGF identification and follow-up implementation are insufficient for detecting expected counterparts; however, improvements in these areas may allow for the recovery of follow-up signals with existing instruments. Future advancements in observational technologies and methodologies will be crucial in furthering these studies. | [
{
"version": "v1",
"created": "Mon, 09 Sep 2024 20:30:57 GMT"
},
{
"version": "v2",
"created": "Mon, 16 Sep 2024 19:49:11 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 16:03:41 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Trigg",
"Aaron C.",
""
],
[
"Stewart",
"Rachael",
""
],
[
"Van Kooten",
"Alex",
""
],
[
"Burns",
"Eric",
""
],
[
"Baring",
"Matthew G.",
""
],
[
"Frederiks",
"Dmitry D.",
""
],
[
"Huppenkothen",
"Daniela",
""
],
[
"O'Connor",
"Brendan",
""
],
[
"Roberts",
"Oliver J.",
""
],
[
"Wadiasingh",
"Zorawar",
""
],
[
"Younes",
"George",
""
],
[
"Bhat",
"Narayana",
""
],
[
"Briggs",
"Michael S.",
""
],
[
"Busmann",
"Malte",
""
],
[
"Goldstein",
"Adam",
""
],
[
"Gruen",
"Daniel",
""
],
[
"Hu",
"Lei",
""
],
[
"Kouveliotou",
"Chryssa",
""
],
[
"Negro",
"Michela",
""
],
[
"Palmese",
"Antonella",
""
],
[
"Riffeser",
"Arno",
""
],
[
"Scotton",
"Lorenzo",
""
],
[
"Svinkin",
"Dmitry S.",
""
],
[
"Veres",
"Peter",
""
],
[
"Zöller",
"Raphael",
""
]
] |
2409.06277 | Yao Shu | Yao Shu, Wenyang Hu, See-Kiong Ng, Bryan Kian Hsiang Low, Fei Richard Yu | Ferret: Federated Full-Parameter Tuning at Scale for Large Language Models | Published as a conference paper at ICML 2025 | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have become indispensable in numerous real-world applications. However, fine-tuning these models at scale, especially in federated settings where data privacy and communication efficiency are critical, presents significant challenges. Existing approaches often resort to parameter-efficient fine-tuning (PEFT) to mitigate communication overhead, but this typically comes at the cost of model accuracy. To this end, we propose federated full-parameter tuning at scale for LLMs (Ferret), the first first-order method with shared randomness to enable scalable full-parameter tuning of LLMs across decentralized data sources while maintaining competitive model accuracy. Ferret accomplishes this through three aspects: (i) it employs widely used first-order methods for efficient local updates; (ii) it projects these updates into a low-dimensional space to considerably reduce communication overhead; and (iii) it reconstructs local updates from this low-dimensional space with shared randomness to facilitate effective full-parameter global aggregation, ensuring fast convergence and competitive final performance. Our rigorous theoretical analyses and insights along with extensive experiments, show that Ferret significantly enhances the scalability of existing federated full-parameter tuning approaches by achieving high computational efficiency, reduced communication overhead, and fast convergence, all while maintaining competitive model accuracy. Our implementation is available at https://github.com/allen4747/Ferret. | [
{
"version": "v1",
"created": "Tue, 10 Sep 2024 07:28:13 GMT"
},
{
"version": "v2",
"created": "Wed, 11 Sep 2024 01:47:48 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 13:15:47 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Shu",
"Yao",
""
],
[
"Hu",
"Wenyang",
""
],
[
"Ng",
"See-Kiong",
""
],
[
"Low",
"Bryan Kian Hsiang",
""
],
[
"Yu",
"Fei Richard",
""
]
] |
2409.06291 | Albert Salmi | Albert Salmi, Miloslav Capek, Lukas Jelinek, Anu Lehtovuori, Ville Viikari | Optimization of Embedded Element Patterns of Reactively Loaded Antenna Arrays | 15 pages | null | null | null | physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | This paper introduces a framework for synthesizing reactively loaded antennas and antenna arrays. The framework comprises two main components: computing the fundamental bound using the semi-definite relaxation and finding a realizable solution via optimization on a Riemannian manifold. The embedded element patterns are subject to the optimization with two distinct goals under study: focusing the radiation in a single direction or synthesizing patterns with desired shapes. The reactive terminations of passive antenna elements serve as optimization variables. We demonstrate the framework using a connected bowtie-slot antenna and antenna array with both beam-focusing and beam-shaping targets. The tests show that the optimization on the Riemannian manifold yields superior results compared to existing methods, such as the genetic algorithm. This is particularly evident in the most complex and extensive problem, which requires the synthesis of shaped embedded element patterns for a sparse reactively loaded antenna array with a limited field of view. | [
{
"version": "v1",
"created": "Tue, 10 Sep 2024 07:44:01 GMT"
},
{
"version": "v2",
"created": "Fri, 13 Dec 2024 12:58:26 GMT"
},
{
"version": "v3",
"created": "Tue, 17 Dec 2024 09:46:39 GMT"
},
{
"version": "v4",
"created": "Mon, 24 Mar 2025 13:59:56 GMT"
},
{
"version": "v5",
"created": "Sun, 08 Jun 2025 12:09:04 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Salmi",
"Albert",
""
],
[
"Capek",
"Miloslav",
""
],
[
"Jelinek",
"Lukas",
""
],
[
"Lehtovuori",
"Anu",
""
],
[
"Viikari",
"Ville",
""
]
] |
2409.06957 | Wei Shen | Chuheng Zhang, Wei Shen, Li Zhao, Xuyun Zhang, Xiaolong Xu, Wanchun Dou, Jiang Bian | Policy Filtration for RLHF to Mitigate Noise in Reward Models | ICML2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | While direct policy optimization methods exist, pioneering LLMs are fine-tuned with reinforcement learning from human feedback (RLHF) to generate better responses under the supervision of a reward model learned from preference data. One major challenge of RLHF is the inaccuracy of the intermediate reward model, especially in the tasks that requires complex reasoning for the reward model to score a response. We find that the reliability of the reward model varies across responses assigned with different rewards. This motivates us to filter the samples whose rewards may be unreliable to improve the signal-to-noise ratio during policy learning, resulting in Policy Filtration for Proximal Policy Optimization (PF-PPO). To choose a proper policy filtering strategy, we use the coefficient of determination (R2) between the rewards and actual scores on filtered samples as the metrics to help us find promising strategies since it measures how well the rewards filtered by PF-PPO indicate real performance. We provide extensive experiments to validate the effectiveness of PF-PPO in code generation and math reasoning tasks. In code generation, PF-PPO achieves the state-of-the-art performance of 7-billion-parameter models on HumanEval (+7.9%), MBPP (+0.7%), and LeetCode Contest (+10.0%) which is a more challenging benchmark created by us. In math reasoning, PF-PPO yields performance increase using different reward models and benchmarks (Ape210K and CMATH). Code is available on https://github.com/DtYXs/verl/tree/pf-ppo. | [
{
"version": "v1",
"created": "Wed, 11 Sep 2024 02:40:38 GMT"
},
{
"version": "v2",
"created": "Tue, 10 Dec 2024 06:21:47 GMT"
},
{
"version": "v3",
"created": "Mon, 26 May 2025 12:34:31 GMT"
},
{
"version": "v4",
"created": "Thu, 29 May 2025 06:23:39 GMT"
},
{
"version": "v5",
"created": "Sat, 07 Jun 2025 02:45:11 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhang",
"Chuheng",
""
],
[
"Shen",
"Wei",
""
],
[
"Zhao",
"Li",
""
],
[
"Zhang",
"Xuyun",
""
],
[
"Xu",
"Xiaolong",
""
],
[
"Dou",
"Wanchun",
""
],
[
"Bian",
"Jiang",
""
]
] |
2409.06985 | Qiushui Xu | Wenhao Zhao, Qiushui Xu, Linjie Xu, Lei Song, Jinyu Wang, Chunlai Zhou, Jiang Bian | Unveiling Markov Heads in Pretrained Language Models for Offline Reinforcement Learning | The first two authors contributed equally. The order was decided by a dice roll | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, incorporating knowledge from pretrained language models (PLMs) into decision transformers (DTs) has generated significant attention in offline reinforcement learning (RL). These PLMs perform well in RL tasks, raising an intriguing question: what kind of knowledge from PLMs has been transferred to RL to achieve such good results? This work first dives into this problem by analyzing each head quantitatively and points out Markov head, a crucial component that exists in the attention heads of PLMs. It leads to extreme attention on the last-input token and performs well only in short-term environments. Furthermore, we prove that this extreme attention cannot be changed by re-training embedding layer or fine-tuning. Inspired by our analysis, we propose a general method GPT2-DTMA, which equips a pretrained DT with Mixture of Attention (MoA), to accommodate diverse attention requirements during fine-tuning. Extensive experiments corroborate our theorems and demonstrate the effectiveness of GPT2-DTMA: it achieves comparable performance in short-term environments while significantly narrowing the performance gap in long-term environments. | [
{
"version": "v1",
"created": "Wed, 11 Sep 2024 03:18:34 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 19:11:05 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Zhao",
"Wenhao",
""
],
[
"Xu",
"Qiushui",
""
],
[
"Xu",
"Linjie",
""
],
[
"Song",
"Lei",
""
],
[
"Wang",
"Jinyu",
""
],
[
"Zhou",
"Chunlai",
""
],
[
"Bian",
"Jiang",
""
]
] |
2409.06990 | Xuzhao Huang | Xuzhao Huang, Akira Seino, Fuyuki Tokuda, Akinari Kobayashi, Dayuan Chen, Yasuhisa Hirata, Norman C. Tien, and Kazuhiro Kosuge | SIS: Seam-Informed Strategy for T-shirt Unfolding | 8 pages, 8 figures. To be published in IEEE Robotics and Automation Letters (RAL) | null | 10.1109/LRA.2025.3574966 | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Seams are information-rich components of garments. The presence of different types of seams and their combinations helps to select grasping points for garment handling. In this paper, we propose a new Seam-Informed Strategy (SIS) for finding actions for handling a garment, such as grasping and unfolding a T-shirt. Candidates for a pair of grasping points for a dual-arm manipulator system are extracted using the proposed Seam Feature Extraction Method (SFEM). A pair of grasping points for the robot system is selected by the proposed Decision Matrix Iteration Method (DMIM). The decision matrix is first computed by multiple human demonstrations and updated by the robot execution results to improve the grasping and unfolding performance of the robot. Note that the proposed scheme is trained on real data without relying on simulation. Experimental results demonstrate the effectiveness of the proposed strategy. The project video is available at https://github.com/lancexz/sis | [
{
"version": "v1",
"created": "Wed, 11 Sep 2024 03:40:39 GMT"
},
{
"version": "v2",
"created": "Thu, 12 Sep 2024 06:56:07 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 07:05:07 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Huang",
"Xuzhao",
""
],
[
"Seino",
"Akira",
""
],
[
"Tokuda",
"Fuyuki",
""
],
[
"Kobayashi",
"Akinari",
""
],
[
"Chen",
"Dayuan",
""
],
[
"Hirata",
"Yasuhisa",
""
],
[
"Tien",
"Norman C.",
""
],
[
"Kosuge",
"Kazuhiro",
""
]
] |
2409.07885 | Ali Emami Kopaei | Ali Emami Kopaei, Karthik Subramaniam Eswaran, Arkadiusz Kosior, Daniel Hodgson, Andrey Matsko, Hossein Taheri, Almut Beige, and Krzysztof Sacha | Towards Timetronics with Photonic Systems | 4 pages + supplemental materials, 3 figures | null | null | null | physics.optics cond-mat.dis-nn cond-mat.quant-gas quant-ph | http://creativecommons.org/licenses/by/4.0/ | Periodic driving of particles can create crystalline structures in their dynamics. Such systems can be used to study solid-state physics phenomena in the time domain. In addition, it is possible to realize photonic time crystals and to engineer the wave-number band structure of optical devices by periodic temporal modulation of the properties of light-propagating media. Here we introduce a versatile approach which uses traveling wave resonators to emulate various condensed matter phases in the time dimension. This is achieved by utilizing temporal modulation of the permittivity and the shape of small segments of the resonators. The required frequency and depth of the modulation are experimentally achievable which opens a pathway for the practical realisation of crystalline structures in time in microwave and in optical systems. | [
{
"version": "v1",
"created": "Thu, 12 Sep 2024 09:44:16 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 08:44:33 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kopaei",
"Ali Emami",
""
],
[
"Eswaran",
"Karthik Subramaniam",
""
],
[
"Kosior",
"Arkadiusz",
""
],
[
"Hodgson",
"Daniel",
""
],
[
"Matsko",
"Andrey",
""
],
[
"Taheri",
"Hossein",
""
],
[
"Beige",
"Almut",
""
],
[
"Sacha",
"Krzysztof",
""
]
] |
2409.08152 | Peter Kazinski | P.O. Kazinski, M.V. Mokrinskiy, and V.A. Ryakin | Surface photoelectric effect by twisted photons as a source of twisted electrons | 19 pp., 2 figs | Proc. R. Soc. A 481, 20240777 (2025) | 10.1098/rspa.2024.0777 | null | cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The theory of surface photoelectric effect by twisted photons is developed. The explicit expression for the probability to record a twisted photoelectron is derived. The conditions when the surface photoelectric effect can be used as a pure source of twisted electrons are found. It is shown that the lightly doped n-InSb crystal with interface without defects at temperatures lower than $2.5$ K satisfies these conditions. The Dirac and Weyl semimetals with electron chemical potential near the top of the Dirac cone obey these conditions at temperatures lower than $60$ K and can also be employed for design of pure sources of twisted electrons by the photoelectric effect. | [
{
"version": "v1",
"created": "Thu, 12 Sep 2024 15:45:34 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kazinski",
"P. O.",
""
],
[
"Mokrinskiy",
"M. V.",
""
],
[
"Ryakin",
"V. A.",
""
]
] |
2409.08469 | Sayan Banerjee | Sayan Banerjee, Krishnakumar Balasubramanian, and Promit Ghosal | Improved Finite-Particle Convergence Rates for Stein Variational Gradient Descent | 26 pages. Some typos corrected in Theorem 3 | null | null | null | math.ST cs.LG math.PR stat.ML stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We provide finite-particle convergence rates for the Stein Variational Gradient Descent (SVGD) algorithm in the Kernelized Stein Discrepancy ($\mathsf{KSD}$) and Wasserstein-2 metrics. Our key insight is that the time derivative of the relative entropy between the joint density of $N$ particle locations and the $N$-fold product target measure, starting from a regular initial distribution, splits into a dominant `negative part' proportional to $N$ times the expected $\mathsf{KSD}^2$ and a smaller `positive part'. This observation leads to $\mathsf{KSD}$ rates of order $1/\sqrt{N}$, in both continuous and discrete time, providing a near optimal (in the sense of matching the corresponding i.i.d. rates) double exponential improvement over the recent result by Shi and Mackey (2024). Under mild assumptions on the kernel and potential, these bounds also grow polynomially in the dimension $d$. By adding a bilinear component to the kernel, the above approach is used to further obtain Wasserstein-2 convergence in continuous time. For the case of `bilinear + Mat\'ern' kernels, we derive Wasserstein-2 rates that exhibit a curse-of-dimensionality similar to the i.i.d. setting. We also obtain marginal convergence and long-time propagation of chaos results for the time-averaged particle laws. | [
{
"version": "v1",
"created": "Fri, 13 Sep 2024 01:49:19 GMT"
},
{
"version": "v2",
"created": "Mon, 30 Sep 2024 01:20:13 GMT"
},
{
"version": "v3",
"created": "Fri, 06 Jun 2025 19:44:32 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Banerjee",
"Sayan",
""
],
[
"Balasubramanian",
"Krishnakumar",
""
],
[
"Ghosal",
"Promit",
""
]
] |
2409.08826 | Wenyi Zhang | Shuqin Pang, Wenyi Zhang | Generalized Nearest Neighbor Decoding: General Input Constellation and a Case Study of Interference Suppression | 14 pages, 6 figures | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, generalized nearest neighbor decoding (GNND), a recently proposed receiver architecture, is studied for channels under general input constellations, and multiuser uplink interference suppression is employed as a case study for demonstrating its potential. In essence, GNND generalizes the well-known nearest neighbor decoding, by introducing a symbol-level memoryless processing step, which can be rendered seamlessly compatible with Gaussian channel-based decoders. First, criteria of the optimal GNND are derived for general input constellations, expressed in the form of conditional moments matching, thereby generalizing the prior work which has been confined to Gaussian input. Then, the optimal GNND is applied to the use case of multiuser uplink, for which the optimal GNND is shown to be capable of achieving information rates nearly identical to the channel mutual information. By contrast, the commonly used channel linearization (CL) approach incurs a noticeable rate loss. A coded modulation scheme is subsequently developed, aiming at implementing GNND using off-the-shelf channel codes, without requiring iterative message passing between demodulator and decoder. Through numerical experiments it is validated that the developed scheme significantly outperforms the CL-based scheme. | [
{
"version": "v1",
"created": "Fri, 13 Sep 2024 13:41:46 GMT"
},
{
"version": "v2",
"created": "Mon, 09 Jun 2025 00:45:24 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Pang",
"Shuqin",
""
],
[
"Zhang",
"Wenyi",
""
]
] |
2409.08906 | Nebiyou Yismaw | Nebiyou Yismaw, Ulugbek S. Kamilov, M. Salman Asif | Gaussian is All You Need: A Unified Framework for Solving Inverse Problems via Diffusion Posterior Sampling | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models can generate a variety of high-quality images by modeling complex data distributions. Trained diffusion models can also be very effective image priors for solving inverse problems. Most of the existing diffusion-based methods integrate data consistency steps by approximating the likelihood function within the diffusion reverse sampling process. In this paper, we show that the existing approximations are either insufficient or computationally inefficient. To address these issues, we propose a unified likelihood approximation method that incorporates a covariance correction term to enhance the performance and avoids propagating gradients through the diffusion model. The correction term, when integrated into the reverse diffusion sampling process, achieves better convergence towards the true data posterior for selected distributions and improves performance on real-world natural image datasets. Furthermore, we present an efficient way to factorize and invert the covariance matrix of the likelihood function for several inverse problems. Our comprehensive experiments demonstrate the effectiveness of our method over several existing approaches. Code available at https://github.com/CSIPlab/CoDPS. | [
{
"version": "v1",
"created": "Fri, 13 Sep 2024 15:20:03 GMT"
},
{
"version": "v2",
"created": "Fri, 06 Jun 2025 23:02:54 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Yismaw",
"Nebiyou",
""
],
[
"Kamilov",
"Ulugbek S.",
""
],
[
"Asif",
"M. Salman",
""
]
] |
2409.08928 | Mathieu Gerber | Yuan Chen, Mathieu Gerber, Christophe Andrieu and Randal Douc | Self-Organizing State-Space Models with Artificial Dynamics | 93 pages (25 pages for the paper and 68 pages for the supplementary material), 9 figures | null | null | null | math.ST stat.CO stat.ME stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the problem of performing parameter and state inference in a state-space model (SSM) parametrized by a static parameter $\theta$. A popular idea to address this problem consists of incorporating $\theta$ in the state of the system and allowing its time evolution, modelled as a Markov chain $(\theta_t)_{t\geq 1}$. This proxy model defines a so-called self-organizing SSM (SO-SSM) to which one may apply standard particle filters. However, the practical implementation of this idea in a theoretically justified manner has remained an open problem until now. In this paper we fill this gap and in particular show that theoretically consistent SO-SSMs can be defined such that $\|\mathrm{Var}(\theta_{t+1}|\theta_{t})\|\rightarrow 0$ slowly as $t\rightarrow\infty$. This, in turn, leads to particle filter algorithms for online inference in SSMs which we find to be robust in simulation. We also develop constructions of $(\theta_t)_{t\geq 1}$ and associated theoretical guarantees tailored to the application of SO-SSMs to maximum likelihood estimation in SSMs, leading to novel iterated filtering algorithms. The algorithms developed in this work have the advantage of being simple to implement and to require minimal tuning to perform well. | [
{
"version": "v1",
"created": "Fri, 13 Sep 2024 15:45:40 GMT"
},
{
"version": "v2",
"created": "Thu, 26 Sep 2024 13:32:39 GMT"
},
{
"version": "v3",
"created": "Thu, 17 Oct 2024 09:35:53 GMT"
},
{
"version": "v4",
"created": "Mon, 11 Nov 2024 09:56:59 GMT"
},
{
"version": "v5",
"created": "Tue, 03 Jun 2025 09:40:44 GMT"
},
{
"version": "v6",
"created": "Mon, 09 Jun 2025 15:20:00 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Chen",
"Yuan",
""
],
[
"Gerber",
"Mathieu",
""
],
[
"Andrieu",
"Christophe",
""
],
[
"Douc",
"Randal",
""
]
] |
2409.09268 | Steven Louis | Steven Louis, Hannah Bradley, Cody Trevillian, Andrei Slavin, Vasyl Tyberkevych | Spintronic Neuron Using a Magnetic Tunnel Junction for Low-Power Neuromorphic Computing | 7 pages, 5 figures, letter | null | null | null | physics.app-ph cond-mat.dis-nn | http://creativecommons.org/licenses/by/4.0/ | This paper proposes a novel spiking artificial neuron design based on a combined spin valve/magnetic tunnel junction (SV/MTJ). Traditional hardware used in artificial intelligence and machine learning faces significant challenges related to high power consumption and scalability. To address these challenges, spintronic neurons, which can mimic biologically inspired neural behaviors, offer a promising solution. We present a model of an SV/MTJ-based neuron which uses technologies that have been successfully integrated with CMOS in commercially available applications. The operational dynamics of the neuron are derived analytically through the Landau-Lifshitz-Gilbert-Slonczewski (LLGS) equation, demonstrating its ability to replicate key spiking characteristics of biological neurons, such as response latency and refractive behavior. Simulation results indicate that the proposed neuron design can operate on a timescale of about 1 ns, without any bias current, and with power consumption as low as 50 uW. | [
{
"version": "v1",
"created": "Sat, 14 Sep 2024 02:27:36 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Oct 2024 14:48:45 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 09:44:07 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Louis",
"Steven",
""
],
[
"Bradley",
"Hannah",
""
],
[
"Trevillian",
"Cody",
""
],
[
"Slavin",
"Andrei",
""
],
[
"Tyberkevych",
"Vasyl",
""
]
] |
2409.09290 | Xunda Sun | Xunda Sun, Xin Wang, Xiangcheng Ma, Kai Wang, Andrew Wetzel, Claude-Andr\'e Faucher-Gigu\`ere, Philip F. Hopkins, Du\v{s}an Kere\v{s}, Russell L. Graf, Andrew Marszewski, Jonathan Stern, Guochao Sun, Lei Sun, Keyer Thyme | The physical origin of positive metallicity radial gradients in high-redshift galaxies: insights from the FIRE-2 cosmological hydrodynamic simulations | 19 pages, 10 figures | null | 10.3847/1538-4357/addab5 | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Using the FIRE-2 cosmological zoom-in simulations, we investigate the temporal evolution of gas-phase metallicity radial gradients of Milky Way-mass progenitors in the redshift range of $0.4<z<3$. We pay special attention to the occurrence of positive (i.e. inverted) metallicity gradients -- where metallicity increases with galactocentric radius. This trend, contrary to the more commonly observed negative radial gradients, has been frequently seen in recent spatially resolved grism observations. The rate of occurrence of positive gradients in FIRE-2 is about $\sim7\%$ for $0.4<z<3$ and $\sim13\%$ at higher redshifts ($1.5<z<3$), broadly consistent with observations. Moreover, we investigate the correlations among galaxy metallicity gradient, stellar mass, star formation rate (SFR), and degree of rotational support. Metallicity gradients show a strong correlation with both sSFR and the rotational-to-dispersion velocity ratio ($v_c/\sigma$), implying that starbursts and kinematic morphology of galaxies play significant roles in shaping these gradients. The FIRE-2 simulations indicate that galaxies with high sSFR (${\rm log(sSFR~[yr^{-1}])}\gtrsim-9.2$) and weak rotational support ($v_c/\sigma\lesssim 1$) are more likely -- by $\sim$15\% -- to develop positive metallicity gradients. This trend is attributed to galaxy-scale gas flows driven by stellar feedback, which effectively redistribute metals within the interstellar medium. Our results support the important role of stellar feedback in governing the chemo-structural evolution and disk formation of Milky Way-mass galaxies at the cosmic noon epoch. | [
{
"version": "v1",
"created": "Sat, 14 Sep 2024 03:40:56 GMT"
},
{
"version": "v2",
"created": "Tue, 06 May 2025 06:38:57 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Sun",
"Xunda",
""
],
[
"Wang",
"Xin",
""
],
[
"Ma",
"Xiangcheng",
""
],
[
"Wang",
"Kai",
""
],
[
"Wetzel",
"Andrew",
""
],
[
"Faucher-Giguère",
"Claude-André",
""
],
[
"Hopkins",
"Philip F.",
""
],
[
"Kereš",
"Dušan",
""
],
[
"Graf",
"Russell L.",
""
],
[
"Marszewski",
"Andrew",
""
],
[
"Stern",
"Jonathan",
""
],
[
"Sun",
"Guochao",
""
],
[
"Sun",
"Lei",
""
],
[
"Thyme",
"Keyer",
""
]
] |
2409.09685 | Marius Lemm | Marius Lemm and Angelo Lucia | On the critical finite-size gap scaling for frustration-free Hamiltonians | 17 pages; final version | null | 10.1142/S0129055X25500151 | null | math-ph math.MP quant-ph | http://creativecommons.org/licenses/by/4.0/ | We prove that the critical finite-size gap scaling for frustration-free Hamiltonians is of inverse-square type. The result covers general graphs embedded in $\mathbb R^D$ and general finite-range interactions without requiring assumptions about the ground state correlations. Therefore, the inverse-square critical gap scaling is a robust, universal property of finite-range frustration-free Hamiltonians. This places further limits on their ability to produce conformal field theories in the continuum limit. Our proof refines the divide-and-conquer strategy of Kastoryano and the second author through the refined Detectability Lemma of Gosset--Huang. | [
{
"version": "v1",
"created": "Sun, 15 Sep 2024 10:37:38 GMT"
},
{
"version": "v2",
"created": "Sat, 07 Jun 2025 11:00:52 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Lemm",
"Marius",
""
],
[
"Lucia",
"Angelo",
""
]
] |
2409.10283 | Sourav Sanyal | Sourav Sanyal and Kaushik Roy | ASMA: An Adaptive Safety Margin Algorithm for Vision-Language Drone Navigation via Scene-Aware Control Barrier Functions | null | null | null | null | cs.RO cs.AI cs.SY eess.IV eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the rapidly evolving field of vision-language navigation (VLN), ensuring safety for physical agents remains an open challenge. For a human-in-the-loop language-operated drone to navigate safely, it must understand natural language commands, perceive the environment, and simultaneously avoid hazards in real time. Control Barrier Functions (CBFs) are formal methods that enforce safe operating conditions. Model Predictive Control (MPC) is an optimization framework that plans a sequence of future actions over a prediction horizon, ensuring smooth trajectory tracking while obeying constraints. In this work, we consider a VLN-operated drone platform and enhance its safety by formulating a novel scene-aware CBF that leverages ego-centric observations from a camera which has both Red-Green-Blue as well as Depth (RGB-D) channels. A CBF-less baseline system uses a Vision-Language Encoder with cross-modal attention to convert commands into an ordered sequence of landmarks. An object detection model identifies and verifies these landmarks in the captured images to generate a planned path. To further enhance safety, an Adaptive Safety Margin Algorithm (ASMA) is proposed. ASMA tracks moving objects and performs scene-aware CBF evaluation on-the-fly, which serves as an additional constraint within the MPC framework. By continuously identifying potentially risky observations, the system performs prediction in real time about unsafe conditions and proactively adjusts its control actions to maintain safe navigation throughout the trajectory. Deployed on a Parrot Bebop2 quadrotor in the Gazebo environment using the Robot Operating System (ROS), ASMA achieves 64%-67% increase in success rates with only a slight increase (1.4%-5.8%) in trajectory lengths compared to the baseline CBF-less VLN. | [
{
"version": "v1",
"created": "Mon, 16 Sep 2024 13:44:50 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Mar 2025 21:51:26 GMT"
},
{
"version": "v3",
"created": "Sat, 07 Jun 2025 18:26:59 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Sanyal",
"Sourav",
""
],
[
"Roy",
"Kaushik",
""
]
] |
2409.10773 | Cedar Site Bai | Cedar Site Bai and Brian Bullins | Tight Lower Bounds under Asymmetric High-Order H\"older Smoothness and Uniform Convexity | ICLR 2025 Oral | ICLR 2025: https://openreview.net/forum?id=fMTPkDEhLQ | null | null | math.OC stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we provide tight lower bounds for the oracle complexity of minimizing high-order H\"older smooth and uniformly convex functions. Specifically, for a function whose $p^{th}$-order derivatives are H\"older continuous with degree $\nu$ and parameter $H$, and that is uniformly convex with degree $q$ and parameter $\sigma$, we focus on two asymmetric cases: (1) $q > p + \nu$, and (2) $q < p+\nu$. Given up to $p^{th}$-order oracle access, we establish worst-case oracle complexities of $\Omega\left( \left( \frac{H}{\sigma}\right)^\frac{2}{3(p+\nu)-2}\left( \frac{\sigma}{\epsilon}\right)^\frac{2(q-p-\nu)}{q(3(p+\nu)-2)}\right)$ in the first case with an $\ell_\infty$-ball-truncated-Gaussian smoothed hard function and $\Omega\left(\left(\frac{H}{\sigma}\right)^\frac{2}{3(p+\nu)-2}+ \log\log\left(\left(\frac{\sigma^{p+\nu}}{H^q}\right)^\frac{1}{p+\nu-q}\frac{1}{\epsilon}\right)\right)$ in the second case, for reaching an $\epsilon$-approximate solution in terms of the optimality gap. Our analysis generalizes previous lower bounds for functions under first- and second-order smoothness as well as those for uniformly convex functions, and furthermore our results match the corresponding upper bounds in this general setting. | [
{
"version": "v1",
"created": "Mon, 16 Sep 2024 23:17:33 GMT"
},
{
"version": "v2",
"created": "Tue, 01 Oct 2024 15:57:06 GMT"
},
{
"version": "v3",
"created": "Sun, 08 Jun 2025 22:42:14 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Bai",
"Cedar Site",
""
],
[
"Bullins",
"Brian",
""
]
] |
2409.10879 | Rawat Kanishka Dr. | R. Kanishka | Primary Ionization and Particle Identification with Straw Tube Detectors | 27 pages, 20 figures | null | null | null | physics.ins-det hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The charged particles are tracked in the high energy physics detectors to provide the information of their properties. One of the tracking detector is straw tube detector that has been used by many experiments. The motivation behind the current work is to study the primary ionization and particle identification using straw tube detectors. Additionally, we report the decay of $^{60}$CO for the study of gamma peaks since these are used in cobalt therapy, that is beneficial for cancer/tumor treatment. The various studies like primary ionization, spatial co-ordinate distributions in the different gas mixtures, transition radiation, drift velocities of electrons and diffusion coefficients using different xenon-based gas mixtures have been obtained. These studies have been done after the optimization of xenon-based gas mixtures for a deeper understanding. The gas mixture that shows maximum transition radiation among the xenon-based gas mixtures was found to be $Xe:He:CH_{4}$ :: 30:55:15. The gas mixture that posses maximum primary ionization has been observed to be $Xe:CO_{2}$ :: 70:30. Finally, the simulations have been carried out for the particle identification in the straw tube detectors with different particles i.e., muons, pions and kaons. | [
{
"version": "v1",
"created": "Tue, 17 Sep 2024 04:20:40 GMT"
},
{
"version": "v2",
"created": "Tue, 22 Apr 2025 15:52:42 GMT"
},
{
"version": "v3",
"created": "Fri, 06 Jun 2025 13:49:57 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Kanishka",
"R.",
""
]
] |
2409.11064 | Jintao Zhang | Mingyue Cheng and Jintao Zhang and Zhiding Liu and Chunli Liu | A Hybrid Multi-Factor Network with Dynamic Sequence Modeling for Early Warning of Intraoperative Hypotension | null | null | null | null | cs.LG cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Intraoperative hypotension (IOH) prediction using past physiological signals is crucial, as IOH may lead to inadequate organ perfusion and significantly elevate the risk of severe complications and mortality. However, current methods often rely on static modeling, overlooking the complex temporal dependencies and the inherently non-stationary nature of physiological signals. We propose a Hybrid Multi-Factor (HMF) network that formulates IOH prediction as a dynamic sequence forecasting task, explicitly capturing both temporal dependencies and physiological non-stationarity. We represent signal dynamics as multivariate time series and decompose them into trend and seasonal components, enabling separate modeling of long-term and periodic variations. Each component is encoded with a patch-based Transformer to balance computational efficiency and feature representation. To address distributional drift from evolving signals, we introduce a symmetric normalization mechanism. Experiments on both public and real-world clinical datasets show that HMF significantly outperforms competitive baselines. We hope HMF offers new insights into IOH prediction and ultimately promotes safer surgical care. Our code is available at https://github.com/Mingyue-Cheng/HMF. | [
{
"version": "v1",
"created": "Tue, 17 Sep 2024 10:46:41 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Feb 2025 07:06:38 GMT"
},
{
"version": "v3",
"created": "Wed, 28 May 2025 08:04:12 GMT"
},
{
"version": "v4",
"created": "Sun, 08 Jun 2025 12:45:29 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Cheng",
"Mingyue",
""
],
[
"Zhang",
"Jintao",
""
],
[
"Liu",
"Zhiding",
""
],
[
"Liu",
"Chunli",
""
]
] |
2409.11554 | Anwar Said | Anwar Said, Waseem Abbas, Xenofon Koutsoukos | PropEnc: A Property Encoder for Graph Neural Networks | conference paper | null | null | null | cs.LG cs.SI | http://creativecommons.org/licenses/by/4.0/ | Graph machine learning, particularly using graph neural networks, heavily relies on node features. However, many real-world systems, such as social and biological networks, lack node features due to privacy concerns, incomplete data, or collection limitations. Structural and positional encoding are commonly used to address this but are constrained by the maximum values of the encoded properties, such as the highest node degree. This limitation makes them impractical for scale-free networks and applications involving large or non-categorical properties. This paper introduces PropEnc, a novel and versatile encoder to generate expressive node embedding from any graph metric. By combining histogram construction with reversed index encoding, PropEnc offers a flexible solution that supports low-dimensional representations and diverse input types, effectively mitigating sparsity issues while improving computational efficiency. Additionally, it replicates one-hot encoding or approximates indices with high accuracy, making it adaptable to a wide range of graph applications. We validate PropEnc through extensive experiments on graph classification task across several social networks lacking node features. The empirical results demonstrate that PropEnc offers an efficient mechanism for constructing node features from various graph metrics. | [
{
"version": "v1",
"created": "Tue, 17 Sep 2024 20:52:50 GMT"
},
{
"version": "v2",
"created": "Sun, 12 Jan 2025 18:15:17 GMT"
},
{
"version": "v3",
"created": "Mon, 09 Jun 2025 15:26:59 GMT"
}
] | 2025-06-10T00:00:00 | [
[
"Said",
"Anwar",
""
],
[
"Abbas",
"Waseem",
""
],
[
"Koutsoukos",
"Xenofon",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.