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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.03407 | Lukas Meyer | Lukas Meyer, Josef Gr\"un, Maximilian Weiherer, Bernhard Egger, Marc Stamminger, Linus Franke | Multi-Spectral Gaussian Splatting with Neural Color Representation | null | null | null | null | cs.GR cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | We present MS-Splatting -- a multi-spectral 3D Gaussian Splatting (3DGS) framework that is able to generate multi-view consistent novel views from images of multiple, independent cameras with different spectral domains. In contrast to previous approaches, our method does not require cross-modal camera calibration and is versatile enough to model a variety of different spectra, including thermal and near-infra red, without any algorithmic changes.
Unlike existing 3DGS-based frameworks that treat each modality separately (by optimizing per-channel spherical harmonics) and therefore fail to exploit the underlying spectral and spatial correlations, our method leverages a novel neural color representation that encodes multi-spectral information into a learned, compact, per-splat feature embedding. A shallow multi-layer perceptron (MLP) then decodes this embedding to obtain spectral color values, enabling joint learning of all bands within a unified representation.
Our experiments show that this simple yet effective strategy is able to improve multi-spectral rendering quality, while also leading to improved per-spectra rendering quality over state-of-the-art methods. We demonstrate the effectiveness of this new technique in agricultural applications to render vegetation indices, such as normalized difference vegetation index (NDVI). | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:36:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Meyer",
"Lukas",
""
],
[
"Grün",
"Josef",
""
],
[
"Weiherer",
"Maximilian",
""
],
[
"Egger",
"Bernhard",
""
],
[
"Stamminger",
"Marc",
""
],
[
"Franke",
"Linus",
""
]
] |
2506.03408 | Yi Xu | Yi Xu, Ruining Yang, Yitian Zhang, Yizhou Wang, Jianglin Lu, Mingyuan Zhang, Lili Su, Yun Fu | Trajectory Prediction Meets Large Language Models: A Survey | 16 pages, GitHub: https://github.com/colorfulfuture/Awesome-Trajectory-Motion-Prediction-Papers | null | null | null | cs.CL cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Recent advances in large language models (LLMs) have sparked growing interest in integrating language-driven techniques into trajectory prediction. By leveraging their semantic and reasoning capabilities, LLMs are reshaping how autonomous systems perceive, model, and predict trajectories. This survey provides a comprehensive overview of this emerging field, categorizing recent work into five directions: (1) Trajectory prediction via language modeling paradigms, (2) Direct trajectory prediction with pretrained language models, (3) Language-guided scene understanding for trajectory prediction, (4) Language-driven data generation for trajectory prediction, (5) Language-based reasoning and interpretability for trajectory prediction. For each, we analyze representative methods, highlight core design choices, and identify open challenges. This survey bridges natural language processing and trajectory prediction, offering a unified perspective on how language can enrich trajectory prediction. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:36:56 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xu",
"Yi",
""
],
[
"Yang",
"Ruining",
""
],
[
"Zhang",
"Yitian",
""
],
[
"Wang",
"Yizhou",
""
],
[
"Lu",
"Jianglin",
""
],
[
"Zhang",
"Mingyuan",
""
],
[
"Su",
"Lili",
""
],
[
"Fu",
"Yun",
""
]
] |
2506.03409 | James Petrie PhD | James Petrie and Onni Aarne | Technical Options for Flexible Hardware-Enabled Guarantees | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | Frontier AI models pose increasing risks to public safety and international security, creating a pressing need for AI developers to provide credible guarantees about their development activities without compromising proprietary information. We propose Flexible Hardware-Enabled Guarantees (flexHEG), a system integrated with AI accelerator hardware to enable verifiable claims about compute usage in AI development. The flexHEG system consists of two primary components: an auditable Guarantee Processor that monitors accelerator usage and verifies compliance with specified rules, and a Secure Enclosure that provides physical tamper protection. In this report, we analyze technical implementation options ranging from firmware modifications to custom hardware approaches, with focus on an "Interlock" design that provides the Guarantee Processor direct access to accelerator data paths. Our proposed architecture could support various guarantee types, from basic usage auditing to sophisticated automated verification. This work establishes technical foundations for hardware-based AI governance mechanisms that could be deployed by 2027 to address emerging regulatory and international security needs in frontier AI development. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:37:43 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Petrie",
"James",
""
],
[
"Aarne",
"Onni",
""
]
] |
2506.03410 | Jared Jonas | Jared Jonas, Bassam Bamieh | An iterative tangential interpolation framework for model reduction of MIMO systems | 13 pages, 4 figures Submitted to IEEE TAC | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | We consider model reduction of large-scale MIMO systems using tangential interpolation in the frequency domain. Our scheme is related to the recently-developed Adaptive Antoulas--Anderson (AAA) algorithm, which is an iterative algorithm that uses concepts from the Loewner framework. Our algorithm uses low-rank interpolation and iteratively adds interpolation points based on several criteria including minimizing maximum errors. We show there is freedom in the interpolation point selection method, leading to multiple algorithms that have trade-offs between computational complexity and approximation performance. We prove that a weighted \(H_2\) norm of a representative error system is monotonically decreasing as interpolation points are added. Finally, we provide computational results and some comparisons with prior works, demonstrating performance on par with standard model reduction methods. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:37:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jonas",
"Jared",
""
],
[
"Bamieh",
"Bassam",
""
]
] |
2506.03411 | Melissa Dutz | Melissa Dutz, Han Shao, Avrim Blum, Aloni Cohen | A Machine Learning Theory Perspective on Strategic Litigation | null | null | null | null | cs.LG cs.GT | http://creativecommons.org/licenses/by/4.0/ | Strategic litigation involves bringing a legal case to court with the goal of having a broader impact beyond resolving the case itself: for example, creating precedent which will influence future rulings. In this paper, we explore strategic litigation from the perspective of machine learning theory. We consider an abstract model of a common-law legal system where a lower court decides new cases by applying a decision rule learned from a higher court's past rulings. In this model, we explore the power of a strategic litigator, who strategically brings cases to the higher court to influence the learned decision rule, thereby affecting future cases. We explore questions including: What impact can a strategic litigator have? Which cases should a strategic litigator bring to court? Does it ever make sense for a strategic litigator to bring a case when they are sure the court will rule against them? | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:38:43 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dutz",
"Melissa",
""
],
[
"Shao",
"Han",
""
],
[
"Blum",
"Avrim",
""
],
[
"Cohen",
"Aloni",
""
]
] |
2506.03412 | Paul Huet | Paul Huet, Quentin Kral and Tristan Guillot | Late gas released in the young Kuiper belt could have significantly contributed to the carbon enrichment of the atmospheres of Neptune and Uranus | null | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | Exo-Kuiper belts have been observed for decades, but the recent detection of gas in some of them may change our view of the Solar System's youth. Late gas produced by the sublimation of CO (or CO$_2$) ices after the dissipation of the primordial gas could be the norm in young planetesimal belts. Hence, a gas-rich Kuiper belt could have been present in the Solar System. The high C/H ratios observed on Uranus and Neptune could be a clue to the existence of such late gas that could have been accreted onto young icy giants. The aim of this paper is to estimate the carbon enrichment of the atmospheres of Uranus and Neptune caused by the accretion of the gas released from a putative gas-rich Kuiper belt. We find that assuming a primordial Kuiper belt with a mass of tens of earth masses leads to significant CO gas accretion onto the giants, which can lead to high C/H ratios, especially for Uranus and Neptune. Our model shows that a relatively massive gas-rich Kuiper belt could have existed in the Solar System's youth, significantly enriching the atmospheres of Uranus and Neptune with carbon. Late gas accretion and its effect on outer giant planets metallicities could be a universal scenario, also occurring in extrasolar systems. Observations of sub-Jupiter exoplanets could provide very useful information to better constrain this scenario, with an enrichment in carbon and oxygen (for warm-enough planets) compared to other elements that should be inversely proportional to their envelope mass. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:43:40 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Huet",
"Paul",
""
],
[
"Kral",
"Quentin",
""
],
[
"Guillot",
"Tristan",
""
]
] |
2506.03414 | Ege Rubak | Adrian Baddeley, Ege Rubak, Suman Rakshit, Gopalan Nair | ROC Curves for Spatial Point Patterns and Presence-Absence Data | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Receiver Operating Characteristic (ROC) curves have recently been used to evaluate the performance of models for spatial presence-absence or presence-only data. Applications include species distribution modelling and mineral prospectivity analysis. We clarify the interpretation of the ROC curve in this context. Contrary to statements in the literature, ROC does not measure goodness-of-fit of a spatial model, and its interpretation as a measure of predictive ability is weak; it is a measure of ranking ability, insensitive to the precise form of the model. To gain insight we draw connections between ROC and existing statistical techniques for spatial point pattern data. The area under the ROC curve (AUC) is related to hypothesis tests of the null hypothesis that the explanatory variables have no effect. The shape of the ROC curve has a diagnostic interpretation. This suggests several new techniques, which extend the scope of application of ROC curves for spatial data, to support variable selection and model selection, analysis of segregation between different types of points, adjustment for a baseline, and analysis of spatial case-control data. The new techniques are illustrated with several real example datasets. Open source R code implementing the techniques is available in the development version of our package spatstat [Baddeley and Turner, 2005, Baddeley et al., 2015] and will be included in the next public release. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:46:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Baddeley",
"Adrian",
""
],
[
"Rubak",
"Ege",
""
],
[
"Rakshit",
"Suman",
""
],
[
"Nair",
"Gopalan",
""
]
] |
2506.03415 | Jay Chan | Jay Chan, Brandon Wang and Paolo Calafiura | Physics and Computing Performance of the EggNet Tracking Pipeline | 7 pages, 5 figures | null | null | null | physics.data-an hep-ex hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Particle track reconstruction is traditionally computationally challenging due to the combinatorial nature of the tracking algorithms employed. Recent developments have focused on novel algorithms with graph neural networks (GNNs), which can improve scalability. While most of these GNN-based methods require an input graph to be constructed before performing message passing, a one-shot approach called EggNet that directly takes detector spacepoints as inputs and iteratively apply graph attention networks with an evolving graph structure has been proposed. The graphs are gradually updated to improve the edge efficiency and purity, thus providing a better model performance. In this work, we evaluate the physics and computing performance of the EggNet tracking pipeline on the full TrackML dataset. We also explore different techniques to reduce constraints on computation memory and computing time. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:48:29 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chan",
"Jay",
""
],
[
"Wang",
"Brandon",
""
],
[
"Calafiura",
"Paolo",
""
]
] |
2506.03416 | Iosif Galanakis | E. Sas{\i}oglu, W. Beida, S. Ghosh, M. Tas, B. Sanyal, S. Lounis, S. Blugel, I. Mertig, and I. Galanakis | Itinerant versus localized magnetism in spin gapped metallic half-Heusler compounds: Stoner criterion and magnetic interactions | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Spin gapped metals have recently emerged as promising candidates for spintronic and nanoelectronic applications, enabling functionalities such as sub-60mV/dec switching, negative differential resistance, and non-local spin-valve effects in field-effect transistors. Realizing these functionalities, however, requires a deeper understanding of their magnetic behavior, which is governed by a subtle interplay between localized and itinerant magnetism. This interplay is particularly complex in spin gapped metallic half-Heusler compounds, whose magnetic properties remain largely unexplored despite previous studies of their electronic structure. In this work, we systematically investigate the magnetic behavior of spin gapped metallic half-Heusler compounds XYZ (X = Fe, Co, Ni, Rh, Ir, Pd, Pt; Y = Ti, V, Zr, Hf, Nb, Ta; Z = In, Sn, Sb), revealing clear trends. Co- and Ni-based compounds predominantly exhibit itinerant magnetism, whereas Ti-, V-, and Fe-based systems may host localized moments, itinerant moments, or a coexistence of both. To uncover the origin of magnetism, we apply the Stoner model, with the Stoner parameter I estimated from Coulomb interaction parameters (Hubbard U and Hund's exchange J) computed using the constrained random phase approximation (cRPA). Our analysis shows that compounds not satisfying the Stoner criterion tend to remain non-magnetic. On the contrary compounds, which satisfy the Stoner criterion, generally exhibit magnetic ordering highlighting the crucial role of electronic correlations and band structure effects in the emergence of magnetism. For compounds with magnetic ground states, we compute Heisenberg exchange parameters, estimate Curie temperatures (T_C), and analyze spin-wave properties, including magnon dispersions and stiffness constants. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:53:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sasıoglu",
"E.",
""
],
[
"Beida",
"W.",
""
],
[
"Ghosh",
"S.",
""
],
[
"Tas",
"M.",
""
],
[
"Sanyal",
"B.",
""
],
[
"Lounis",
"S.",
""
],
[
"Blugel",
"S.",
""
],
[
"Mertig",
"I.",
""
],
[
"Galanakis",
"I.",
""
]
] |
2506.03417 | Xuwen Zhang | Guofang Wang, Wei Wei, Xuwen Zhang | Half-space Liouville-type theorems for minimal graphs with capillary boundary | null | null | null | null | math.DG math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we prove two Liouville-type theorems for capillary minimal graph over $\mathbb{R}^n_+$. First, if $u$ has linear growth, then for $n=2,3$ and for any $\theta\in(0,\pi)$, or $n\geq4$ and $\theta\in(\frac{\pi}6,\frac{5\pi}6)$, $u$ must be flat. Second, if $u$ is one-sided bounded on $\mathbb{R}^n_+$, then for any $n$ and $\theta\in(0,\pi)$, $u$ must be flat. The proofs build upon gradient estimates for the mean curvature equation over $\mathbb{R}^n_+$ with capillary boundary condition, which are based on carefully adapting the maximum principle to the capillary setting. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:54:01 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Guofang",
""
],
[
"Wei",
"Wei",
""
],
[
"Zhang",
"Xuwen",
""
]
] |
2506.03418 | Cory Palmer | D\'aniel Gerbner and Cory Palmer | Survey of generalized Tur\'an problems -- counting subgraphs | We welcome comments from the community, including suggested references. We especially encourage you to share any generalized Tur\'an open problems to be included in Section 6! | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For fixed graphs $H$ and $F$, the \emph{generalized Tur\'an number} $\mathrm{ex}(n,H,F)$ is the maximum possible number of copies of a subgraph $H$ in an $n$-vertex $F$-free graph. This article is a survey of this extremal function whose study was initiated in an influential 2016 article by Alon and Shikhelman (\emph{J. Combin. Theory, B}, {\bf 121}, 2016). | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:57:06 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gerbner",
"Dániel",
""
],
[
"Palmer",
"Cory",
""
]
] |
2506.03419 | Pablo Groisman | Pablo Groisman, Cecilia De Vita, Juli\'an Fern\'andez Bonder, Yuanzhao Zhang | The size of the sync basin resolved | null | null | null | null | math-ph math.DS math.MP nlin.AO nlin.PS | http://creativecommons.org/licenses/by/4.0/ | Sparsely coupled Kuramoto oscillators offer a fertile playground for exploring high-dimensional basins of attraction due to their simple yet multistable dynamics. For $n$ identical Kuramoto oscillators on cycle graphs, it is well known that the only attractors are twisted states, whose phases wind around the circle with a constant gap between neighboring oscillators ($\theta_j = 2\pi q j/n$). It was conjectured in 2006 that basin sizes of these twisted states scale as $e^{-kq^2}$ to the winding number $q$. Here, we provide new numerical and analytical evidence supporting the conjecture and uncover the dynamical mechanism behind the Gaussian scaling. The key idea is that, when starting with a random initial condition, the winding number of the solution stabilizes rapidly at $t \propto \log n$, before long-range correlation can develop among oscillators. This timescale separation allows us to model the winding number as a sum of weakly dependent variables, leading to a Central Limit Theorem derivation of the basin scaling. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 21:57:18 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Groisman",
"Pablo",
""
],
[
"De Vita",
"Cecilia",
""
],
[
"Bonder",
"Julián Fernández",
""
],
[
"Zhang",
"Yuanzhao",
""
]
] |
2506.03420 | Muhammad Zubair Hasan | Muhammad Zubair Hasan and Fahmida Yasmin Rifat | Hybrid Ensemble of Segmentation-Assisted Classification and GBDT for Skin Cancer Detection with Engineered Metadata and Synthetic Lesions from ISIC 2024 Non-Dermoscopic 3D-TBP Images | Written as per the requirements of CVPR 2025. It is a 8 page paper without reference | null | null | null | eess.IV cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Skin cancer is among the most prevalent and life-threatening diseases worldwide, with early detection being critical to patient outcomes. This work presents a hybrid machine and deep learning-based approach for classifying malignant and benign skin lesions using the SLICE-3D dataset from ISIC 2024, which comprises 401,059 cropped lesion images extracted from 3D Total Body Photography (TBP), emulating non-dermoscopic, smartphone-like conditions. Our method combines vision transformers (EVA02) and our designed convolutional ViT hybrid (EdgeNeXtSAC) to extract robust features, employing a segmentation-assisted classification pipeline to enhance lesion localization. Predictions from these models are fused with a gradient-boosted decision tree (GBDT) ensemble enriched by engineered features and patient-specific relational metrics. To address class imbalance and improve generalization, we augment malignant cases with Stable Diffusion-generated synthetic lesions and apply a diagnosis-informed relabeling strategy to harmonize external datasets into a 3-class format. Using partial AUC (pAUC) above 80 percent true positive rate (TPR) as the evaluation metric, our approach achieves a pAUC of 0.1755 -- the highest among all configurations. These results underscore the potential of hybrid, interpretable AI systems for skin cancer triage in telemedicine and resource-constrained settings. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:00:03 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hasan",
"Muhammad Zubair",
""
],
[
"Rifat",
"Fahmida Yasmin",
""
]
] |
2506.03421 | Hongjun Gao | Shanqi Liu, Yaozhong Hu and Hongjun Gao | Central limit theorem of Multilevel Monte Carlo Euler estimators for Stochastic Volterra equations with fractional kernels | 52 pages | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper is devoted to proving a (Lindeberg-Feller type ) central limit theorem for the multilevel Monte Carlo estimator associated with the Euler discretization scheme for the stochastic Volterra equations with fractional kernels $K(u)=u^{H-\frac{1}{2}}/\Gamma(H+1/2), H\in (0,1/2]$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:00:13 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Liu",
"Shanqi",
""
],
[
"Hu",
"Yaozhong",
""
],
[
"Gao",
"Hongjun",
""
]
] |
2506.03422 | Anton Hinneck | Anton Hinneck, Mathias Duckheim, Michael Metzger, Stefan Niessen | On Loss-Minimal Radial Topologies in MV Systems | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | Distribution system reconfiguration (DSR) means optimizing the topology of a distribution grid using switching actions. Switching actions are a degrees of freedom available to distribution system operators, e.g. to manage planned and unplanned outages. DSR is a NP-hard combinatorial problem. Finding good or even optimal solutions is computationally expensive. While transmission and high-voltage grids are generally operated in a meshed state, MV distribution systems are commonly operated as radial networks even though meshed operation would be supported. This improves resilience because faults can be isolated more easily keeping the rest of the system operational and minimizing impact on customers. We propose an AC DSR formulation and benchmark it against a common formulation from the literature. Our results indicate that additional acyclicity constraints can significantly improve solver performance. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:01:02 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hinneck",
"Anton",
""
],
[
"Duckheim",
"Mathias",
""
],
[
"Metzger",
"Michael",
""
],
[
"Niessen",
"Stefan",
""
]
] |
2506.03423 | Timothy Mahoney | Timothy B Mahoney, David B Grayden, Sam E John | Sub-Scalp EEG for Sensorimotor Brain-Computer Interface | 43 Pages, 9 Figures, 3 Tables | null | null | null | eess.SP q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Objective: To establish sub-scalp electroencephalography (EEG) as a viable option for brain-computer interface (BCI) applications, particularly for chronic use, by demonstrating its effectiveness in recording and classifying sensorimotor neural activity. Approach: Two experiments were conducted in this study. The first aim was to demonstrate the high spatial resolution of sub-scalp EEG through analysis of somatosensory evoked potentials in sheep models. The second focused on the practical application of sub-scalp EEG, classifying motor execution using data collected during a sheep behavioural experiment. Main Results: We successfully demonstrated the recording of sensorimotor rhythms using sub-scalp EEG in sheep models. Important spatial, temporal, and spectral features of these signals were identified, and we were able to classify motor execution with above-chance performance. These results are comparable to previous work that investigated signal quality and motor execution classification using ECoG and endovascular arrays in sheep models. Significance: These results suggest that sub-scalp EEG may provide signal quality that approaches that of more invasive neural recording methods such as ECoG and endovascular arrays, and support the use of sub-scalp EEG for chronic BCI applications. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:03:58 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mahoney",
"Timothy B",
""
],
[
"Grayden",
"David B",
""
],
[
"John",
"Sam E",
""
]
] |
2506.03424 | Hanan Samet | Nicole R Schneider and Nandini Ramachandran and Kent O'Sullivan and Hanan Samet | DistRAG: Towards Distance-Based Spatial Reasoning in LLMs | null | null | null | null | cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | Many real world tasks where Large Language Models (LLMs) can be used require spatial reasoning, like Point of Interest (POI) recommendation and itinerary planning. However, on their own LLMs lack reliable spatial reasoning capabilities, especially about distances. To address this problem, we develop a novel approach, DistRAG, that enables an LLM to retrieve relevant spatial information not explicitly learned during training. Our method encodes the geodesic distances between cities and towns in a graph and retrieves a context subgraph relevant to the question. Using this technique, our method enables an LLM to answer distance-based reasoning questions that it otherwise cannot answer. Given the vast array of possible places an LLM could be asked about, DistRAG offers a flexible first step towards providing a rudimentary `world model' to complement the linguistic knowledge held in LLMs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:10:39 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Schneider",
"Nicole R",
""
],
[
"Ramachandran",
"Nandini",
""
],
[
"O'Sullivan",
"Kent",
""
],
[
"Samet",
"Hanan",
""
]
] |
2506.03425 | Surya Koppisetti | Petr Grinberg, Ankur Kumar, Surya Koppisetti, Gaurav Bharaj | A Data-Driven Diffusion-based Approach for Audio Deepfake Explanations | 5 pages, 3 figures, accepted at Interspeech 2025 | null | null | null | eess.AS cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Evaluating explainability techniques, such as SHAP and LRP, in the context of audio deepfake detection is challenging due to lack of clear ground truth annotations. In the cases when we are able to obtain the ground truth, we find that these methods struggle to provide accurate explanations. In this work, we propose a novel data-driven approach to identify artifact regions in deepfake audio. We consider paired real and vocoded audio, and use the difference in time-frequency representation as the ground-truth explanation. The difference signal then serves as a supervision to train a diffusion model to expose the deepfake artifacts in a given vocoded audio. Experimental results on the VocV4 and LibriSeVoc datasets demonstrate that our method outperforms traditional explainability techniques, both qualitatively and quantitatively. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:10:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Grinberg",
"Petr",
""
],
[
"Kumar",
"Ankur",
""
],
[
"Koppisetti",
"Surya",
""
],
[
"Bharaj",
"Gaurav",
""
]
] |
2506.03426 | Joonseong Kang | Joonseong Kang, Soojeong Lee, Subeen Park, Sumin Park, Taero Kim, Jihee Kim, Ryunyi Lee, Kyungwoo Song | Adaptive Task Vectors for Large Language Models | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | In-Context Learning (ICL) enables Large Language Models (LLMs) to perform tasks without parameter updates by conditioning on a few demonstrations provided in the prompt. Despite its success, ICL suffers from several limitations, including sensitivity to demonstration order, context length constraints, and computational inefficiency. To address these challenges, task vector-based approaches compress task information into a single vector. However, these methods typically construct task vectors from fixed sets of demonstrations and reuse them across input queries, without conditioning on the specific input. This limitation can lead models to struggle with effective adaptation when the input query is not well aligned with the underlying demonstrations, consequently degrading their generalization performance on unseen tasks. To overcome this limitation, we propose Adaptive Task Vectors (ATV), a simple and effective framework that dynamically generates task vectors conditioned on each input query. ATV employs a small language model to generate task vectors, which are then transformed to match the target LLM's architecture and applied to guide its output generation. In contrast to ICL and previous vector-based approaches, which rely on fixed demonstration sets and their corresponding vectors, ATV dynamically generates task vectors tailored to each specific input query and task. Consequently, ATV demonstrates strong performance and generalization capabilities, even for unseen tasks. Furthermore, we provide a theoretical analysis indicating that ATV is expressively equivalent to LoRA under equal rank budgets and more expressive than Prefix-Tuning, thereby offering formal support for its representational advantage. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:12:28 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kang",
"Joonseong",
""
],
[
"Lee",
"Soojeong",
""
],
[
"Park",
"Subeen",
""
],
[
"Park",
"Sumin",
""
],
[
"Kim",
"Taero",
""
],
[
"Kim",
"Jihee",
""
],
[
"Lee",
"Ryunyi",
""
],
[
"Song",
"Kyungwoo",
""
]
] |
2506.03427 | Sungkun Hong | Seyed K. Alavi, Youssef Ezzo, Ashik Pulikkathara, Sungkun Hong | Interference-enhanced light-matter interaction of an optically levitated dielectric particle | null | null | null | null | quant-ph physics.optics | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Enhancing light-matter interaction is a central subject in quantum science. In this work, we investigate an interference-based approach to amplify the effect of a weak optical field on a nanoscale particle levitated in an optical tweezer. By interfering the weak signal field with the strong trapping field, we achieve a significant enhancement of the optical force caused by the weak field, compared to the force in the absence of interference. This enhanced interaction enables us to detect the weak field through monitoring the motion of the particle, with sensitivity down to 79 pW/Hz. We discuss the potential of this approach for realizing an ultrasensitive, nondestructive light field detector and for exploring optomechanical interactions at the single-photon level. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:14:00 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Alavi",
"Seyed K.",
""
],
[
"Ezzo",
"Youssef",
""
],
[
"Pulikkathara",
"Ashik",
""
],
[
"Hong",
"Sungkun",
""
]
] |
2506.03428 | Haoyuan Li | Haoyuan Li, Nan Wang, Leon Zhang, Sanghoon Song, Yanwen Sun, May-Ling Ng, Takahiro Sato, Dillon Hanlon, Sajal Dahal, Mario D. Balcazar, Vincent Esposito, Selene She, Chance Caleb Ornelas-Skarin, Joan Vila-Comamala, Christian David, Nadia Berndt, Peter Richard Miedaner, Zhuquan Zhang, Matthias Ihme, Mariano Trigo, Keith A. Nelson, Jerome B. Hastings, Alexei A. Maznev, Laura Foglia, Samuel Teitelbaum, David A. Reis, Diling Zhu | Nanoscale Ultrafast Lattice Modulation with Hard X-ray Free Electron Laser | 35 pages, 3 figures, 9 extended data figures | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | Understanding and controlling microscopic dynamics across spatial and temporal scales has driven major progress in science and technology over the past several decades. While ultrafast laser-based techniques have enabled probing nanoscale dynamics at their intrinsic temporal scales down to femto- and attoseconds, the long wavelengths of optical lasers have prevented the interrogation and manipulation of such dynamics with nanoscale spatial specificity. With advances in hard X-ray free electron lasers (FELs), significant progress has been made developing X-ray transient grating (XTG) spectroscopy, aiming at the coherent control of elementary excitations with nanoscale X-ray standing waves. So far, XTGs have been probed only at optical wavelengths, thus intrinsically limiting the achievable periodicities to several hundreds of nm. By achieving sub-femtosecond synchronization of two hard X-ray pulses at a controlled crossing angle, we demonstrate the generation of an XTG with spatial periods of 10 nm. The XTG excitation drives a thermal grating that drives coherent monochromatic longitudinal acoustic phonons in the cubic perovskite, SrTiO3 (STO). With a third X-ray pulse with the same photon energy, time-and-momentum resolved measurement of the XTG-induced scattering intensity modulation provides evidence of ballistic thermal transport at nanometer scale in STO. These results highlight the great potential of XTG for studying high-wave-vector excitations and nanoscale transport in condensed matter, and establish XTG as a powerful platform for the coherent control and study of nanoscale dynamics. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:19:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Li",
"Haoyuan",
""
],
[
"Wang",
"Nan",
""
],
[
"Zhang",
"Leon",
""
],
[
"Song",
"Sanghoon",
""
],
[
"Sun",
"Yanwen",
""
],
[
"Ng",
"May-Ling",
""
],
[
"Sato",
"Takahiro",
""
],
[
"Hanlon",
"Dillon",
""
],
[
"Dahal",
"Sajal",
""
],
[
"Balcazar",
"Mario D.",
""
],
[
"Esposito",
"Vincent",
""
],
[
"She",
"Selene",
""
],
[
"Ornelas-Skarin",
"Chance Caleb",
""
],
[
"Vila-Comamala",
"Joan",
""
],
[
"David",
"Christian",
""
],
[
"Berndt",
"Nadia",
""
],
[
"Miedaner",
"Peter Richard",
""
],
[
"Zhang",
"Zhuquan",
""
],
[
"Ihme",
"Matthias",
""
],
[
"Trigo",
"Mariano",
""
],
[
"Nelson",
"Keith A.",
""
],
[
"Hastings",
"Jerome B.",
""
],
[
"Maznev",
"Alexei A.",
""
],
[
"Foglia",
"Laura",
""
],
[
"Teitelbaum",
"Samuel",
""
],
[
"Reis",
"David A.",
""
],
[
"Zhu",
"Diling",
""
]
] |
2506.03429 | Vahid Nourozi | Vahid Nourozi | Flagged Extensions and Numerical Simulations for Quantum Channel Capacity: Bridging Theory and Computation | null | null | null | null | quant-ph cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | I will investigate the capacities of noisy quantum channels through a combined analytical and numerical approach. First, I introduce novel flagged extension techniques that embed a channel into a higher-dimensional space, enabling single-letter upper bounds on quantum and private capacities. My results refine previous bounds and clarify noise thresholds beyond which quantum transmission vanishes. Second, I present a simulation framework that uses coherent information to estimate channel capacities in practice, focusing on two canonical examples: the amplitude damping channel (which we confirm is degradable and thus single-letter) and the depolarizing channel (whose capacity requires multi-letter superadditivity). By parameterizing input qubit states on the Bloch sphere, I numerically pinpoint the maximum coherent information for each channel and validate the flagged extension bounds. Notably, I capture the abrupt transition to zero capacity at high noise and observe superadditivity for moderate noise levels. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:21:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nourozi",
"Vahid",
""
]
] |
2506.03430 | Emmanuel Badmus | Emmanuel O. Badmus, Amritanshu Pandey | Two-Stage Bidirectional Inverter Equivalent Circuit Model for Distribution Grid Steady-State Analysis and Optimization | null | null | null | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a \textit{physics-based} steady-state equivalent circuit model of a two-stage bidirectional inverter. These inverters connect distributed energy resources (DERs), such as photovoltaic (PV) and battery systems, to distribution grids. Existing inverter models have technical gaps on three fronts: i) inadequate modeling of inverter losses, ii) use of mathematical abstractions for bidirectional flow of power, and iii) inability to integrate different control modes into nonlinear solvers without loss of generality. We propose a physics-first model that explicitly captures losses in passive circuit components based on circuit-level principles. We enable bidirectional power flow without binary or complementarity constraints by formulating loss terms as smooth, sign-aware expressions of current. We introduce and parameterize controlled current sources with twice-differentiable continuous functions to enable inverter control modes without loss of generality. We integrate DERs with the proposed inverter model at the load buses of distribution networks to perform power flow and optimization studies on real-world distribution networks with over 20,000 nodes. We demonstrate that the proposed model is more accurate, integrates seamlessly with various control modes without loss of generality, and scales robustly to large optimization problems.
Index Terms: bidirectional inverter model, circuit-based modeling, DERs, inverter efficiency, power control, steady-state analysis. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:22:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Badmus",
"Emmanuel O.",
""
],
[
"Pandey",
"Amritanshu",
""
]
] |
2506.03431 | Josep Gallegos Saliner | Josep M. Gallegos | One-sided Rellich inequalities, Regularity problem and uniform rectifiability | 36 pages, 3 figures | null | null | null | math.AP math.CA | http://creativecommons.org/licenses/by/4.0/ | Let $\Omega\subset \mathbb R^{n+1}$, $n\geq1$, be a bounded open set satisfying the interior corkscrew condition with a uniformly $n$-rectifiable boundary but without any connectivity assumptions. We establish the estimate $$ \Vert \partial_\nu u_f \Vert_{M} \lesssim \Vert \nabla_H f \Vert_{L^1(\partial\Omega)}, \quad \mbox{for all $f\in\operatorname{Lip}(\partial\Omega)$} $$ where $u_f$ is the solution to the Dirichlet problem with boundary data $f$, $\partial_\nu u_f$ is the normal derivative of $u_f$ at the boundary in the weak sense, $\Vert \cdot \Vert_{M}$ denotes the total variation norm and $\nabla_H f$ is the Haj{\l}asz-Sobolev gradient of $f$.
Conversely, if $\Omega\subset \mathbb R^{n+1}$ is a corkscrew domain with $n$-Ahlfors regular boundary and the previous inequality holds for solutions to the Dirichlet problem on $\Omega$, then $\partial\Omega$ must satisfy the weak-no-boxes condition introduced by David and Semmes. Hence, in the planar case, the one-sided Rellich inequality characterizes the uniform rectifiability of $\partial\Omega$.
We also show solvability of the regularity problem in weak $L^1$ for bounded corkscrew domains with a uniformly $n$-rectifiable boundary, that is $$\Vert N(\nabla u_f) \Vert_{L^{1,\infty}(\partial\Omega)} \lesssim \Vert \nabla_H f\Vert_{L^1(\partial\Omega)},\quad \mbox{for all $f\in\operatorname{Lip}(\partial\Omega)$}$$ where $N$ is the nontangential maximal operator. As an application of our results, we prove that for general elliptic operators, the solvability of the Dirichlet problem does not imply the solvability of the regularity problem. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:26:09 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gallegos",
"Josep M.",
""
]
] |
2506.03432 | Luana L. Silva Ribeiro Ph.D. | Cleonice F. Bracciali and Karina S. Rampazzi and Luana L. Silva Ribeiro | Differential equations for a class of semiclassical orthogonal polynomials on the unit circle | null | null | null | null | math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider semiclassical orthogonal polynomials on the unit circle associated with a weight function that satisfy a Pearson-type differential equation involving two polynomials of degree at most three. Structure relations and difference equations for these orthogonal polynomials are found, and, as a consequence, explicit first and second order differential equations are derived. Among the applications, differential equations for a family of polynomials that generalizes the Jacobi polynomials on the unit circle and the modified Bessel polynomials are established. It is also shown that in some cases the Verblunsky coefficients satisfy a discrete Painlev\'e II equation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:31:38 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bracciali",
"Cleonice F.",
""
],
[
"Rampazzi",
"Karina S.",
""
],
[
"Ribeiro",
"Luana L. Silva",
""
]
] |
2506.03433 | Yifan Li | Yifan Li, Xin Li, Tianqin Li, Wenbin He, Yu Kong, Liu Ren | ViT-Split: Unleashing the Power of Vision Foundation Models via Efficient Splitting Heads | The project is available: https://jackyfl.github.io/vitsplit.github.io/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Vision foundation models (VFMs) have demonstrated remarkable performance across a wide range of downstream tasks. While several VFM adapters have shown promising results by leveraging the prior knowledge of VFMs, we identify two inefficiencies in these approaches. First, the interaction between convolutional neural network (CNN) and VFM backbone triggers early layer gradient backpropagation. Second, existing methods require tuning all components, adding complexity. Besides, these adapters alter VFM features, underutilizing the prior knowledge. To tackle these challenges, we propose a new approach called ViT-Split, based on a key observation: the layers of several VFMs, like DINOv2, can be divided into two distinct components: an extractor for learning low-level features and an adapter for learning task-specific features. Leveraging this insight, we eliminate the CNN branch and introduce two heads, task head and prior head, to the frozen VFM. The task head is designed to learn task-specific features, mitigating the early gradient propagation issue. The prior head is used to leverage the multi-scale prior features from the frozen VFM, reducing tuning parameters and overfitting. Extensive experiments on various tasks (e.g., segmentation, detection, depth estimation, and visual question answering) validate the effectiveness and efficiency of ViT-Split. Specifically, ViT-Split reduces training time up to $4\times$ while achieving comparable or even better results on ADE20K, compared to other VFM adapters. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:34:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Li",
"Yifan",
""
],
[
"Li",
"Xin",
""
],
[
"Li",
"Tianqin",
""
],
[
"He",
"Wenbin",
""
],
[
"Kong",
"Yu",
""
],
[
"Ren",
"Liu",
""
]
] |
2506.03434 | Ahmad Dawar Hakimi | Ahmad Dawar Hakimi, Ali Modarressi, Philipp Wicke, Hinrich Sch\"utze | Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Understanding how large language models (LLMs) acquire and store factual knowledge is crucial for enhancing their interpretability and reliability. In this work, we analyze the evolution of factual knowledge representation in the OLMo-7B model by tracking the roles of its attention heads and feed forward networks (FFNs) over the course of pre-training. We classify these components into four roles: general, entity, relation-answer, and fact-answer specific, and examine their stability and transitions. Our results show that LLMs initially depend on broad, general-purpose components, which later specialize as training progresses. Once the model reliably predicts answers, some components are repurposed, suggesting an adaptive learning process. Notably, attention heads display the highest turnover. We also present evidence that FFNs remain more stable throughout training. Furthermore, our probing experiments reveal that location-based relations converge to high accuracy earlier in training than name-based relations, highlighting how task complexity shapes acquisition dynamics. These insights offer a mechanistic view of knowledge formation in LLMs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:35:09 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hakimi",
"Ahmad Dawar",
""
],
[
"Modarressi",
"Ali",
""
],
[
"Wicke",
"Philipp",
""
],
[
"Schütze",
"Hinrich",
""
]
] |
2506.03435 | Brian Barch | Brian Barch, Daniel Lidar | Computational Complexity of Non-Hermitian Quantum Systems | 11 + 3 pages, comments welcome | null | null | null | quant-ph cs.CC | http://creativecommons.org/licenses/by/4.0/ | We analyze the computational power of non-Hermitian quantum dynamics, i.e., conditional time evolutions that arise when a quantum system is monitored and one postselects on a particular measurement record. We establish an approximate equivalence between post-selection and arbitrary non-Hermitian Hamiltonians. Namely, first we establish hardness in the following sense: Let $U=e^{-iHt}$ be an NH gate on $n$ qubits whose smallest and largest singular values differ by at least $2^{-\text{poly}(n)}$. Together with any universal set of unitary gates, the ability to apply such a gate lets one efficiently emulate postselection. The resulting model decides every language in PostBQP; hence, under standard complexity conjectures, fully scalable NH quantum computers are unlikely to be engineered. Second, we establish upper bounds which show that conversely, any non-Hermitian evolution can be written as a unitary on a system-meter pair followed by postselecting the meter. This ``purification'' is compact -- it introduces only $O(\delta^{2})$ Trotter error per time step $\delta$ -- so any NH model whose purification lies in a strongly simulable unitary family (e.g., Clifford, matchgate, or low-bond-dimension tensor-network circuits) remains efficiently simulable. Thus, non-Hermitian physics neither guarantees a quantum advantage nor precludes efficient classical simulation: its complexity is controlled by the singular-value radius of the evolution operator and by the structure of its unitary purification. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:35:28 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Barch",
"Brian",
""
],
[
"Lidar",
"Daniel",
""
]
] |
2506.03436 | Eleanor Byrnes | Eleanor Byrnes, Bernard Deconinck, Anastassiya Semenova | Stokes Waves in Water of Finite Depth | null | null | null | null | nlin.PS | http://creativecommons.org/licenses/by/4.0/ | Periodic water waves of permanent form traveling at constant speed, the so-called Stokes waves, are studied in water of fixed finite depth using methods previously used in water of infinite depth. We apply our methods to waves of varying steepness over a range of fixed depths in order to determine how a number of physical quantities related to the waves change as the steepness of the waves increases. Finally, we examine the Taylor sign condition for these waves, as well as the complex singularities outside of their domain of definition when the waves are considered as a function of a conformal variable. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:37:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Byrnes",
"Eleanor",
""
],
[
"Deconinck",
"Bernard",
""
],
[
"Semenova",
"Anastassiya",
""
]
] |
2506.03437 | Jason Mohoney | Jason Mohoney, Devesh Sarda, Mengze Tang, Shihabur Rahman Chowdhury, Anil Pacaci, Ihab F. Ilyas, Theodoros Rekatsinas, Shivaram Venkataraman | Quake: Adaptive Indexing for Vector Search | null | null | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Vector search, the task of finding the k-nearest neighbors of high-dimensional vectors, underpins many machine learning applications, including recommendation systems and information retrieval. However, existing approximate nearest neighbor (ANN) methods perform poorly under dynamic, skewed workloads where data distributions evolve. We introduce Quake, an adaptive indexing system that maintains low latency and high recall in such environments. Quake employs a hierarchical partitioning scheme that adjusts to updates and changing access patterns, guided by a cost model that predicts query latency based on partition sizes and access frequencies. Quake also dynamically optimizes query execution parameters to meet recall targets using a novel recall estimation model. Furthermore, Quake utilizes optimized query processing, leveraging NUMA-aware parallelism for improved memory bandwidth utilization. To evaluate Quake, we prepare a Wikipedia vector search workload and develop a workload generator to create vector search workloads with configurable access patterns. Our evaluation shows that on dynamic workloads, Quake achieves query latency reductions of 1.5-22x and update latency reductions of 6-83x compared to state-of-the-art indexes SVS, DiskANN, HNSW, and SCANN. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:37:37 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mohoney",
"Jason",
""
],
[
"Sarda",
"Devesh",
""
],
[
"Tang",
"Mengze",
""
],
[
"Chowdhury",
"Shihabur Rahman",
""
],
[
"Pacaci",
"Anil",
""
],
[
"Ilyas",
"Ihab F.",
""
],
[
"Rekatsinas",
"Theodoros",
""
],
[
"Venkataraman",
"Shivaram",
""
]
] |
2506.03438 | Nima Razavi | Nima Razavi, Murat Bayraktar, Nuria Gonzalez Prelcic, Robert W. Heath Jr | Interference-Aware Multiuser Hybrid Precoding for Coexistence with LEO Satellite Communication | Submitted to MILCOM 2025 | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Interference from terrestrial networks can reduce the communication rate for low Earth orbit (LEO) satellites in the upper mid-band. To coexist in frequency, MIMO precoding can be used to reduce the signal that impinges on the LEO satellite. We present a beamforming algorithm designed for the hybrid architecture that incorporates a satellite interference penalty while optimizing the analog and digital precoders. Our algorithm optimizes the precoding at the base station (BS) within the set of precoders that null the interference to the satellite. Simulations demonstrate that our algorithm reduces the interference at the satellites and lowers the probability of violating prescribed LEO satellite protection thresholds, outperforming prior hybrid nulling algorithms. Results indicate that the algorithm maintains sum-rate within 3\% of the existing hybrid solutions, while effectively improving interference to noise power by 22.4 dB. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:42:38 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Razavi",
"Nima",
""
],
[
"Bayraktar",
"Murat",
""
],
[
"Prelcic",
"Nuria Gonzalez",
""
],
[
"Heath",
"Robert W.",
"Jr"
]
] |
2506.03441 | James Sud | Anuj Apte, Ojas Parekh, and James Sud | Conjectured Bounds for 2-Local Hamiltonians via Token Graphs | 40 pages, 2 figures | null | null | null | quant-ph cs.DS | http://creativecommons.org/licenses/by/4.0/ | We explain how the maximum energy of the Quantum MaxCut, XY, and EPR Hamiltonians on a graph $G$ are related to the spectral radii of the token graphs of $G$. From numerical study, we conjecture new bounds for these spectral radii based on properties of $G$. We show how these conjectures tighten the analysis of existing algorithms, implying state-of-the-art approximation ratios for all three Hamiltonians. Our conjectures also provide simple combinatorial bounds on the ground state energy of the antiferromagnetic Heisenberg model, which we prove for bipartite graphs. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:52:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Apte",
"Anuj",
""
],
[
"Parekh",
"Ojas",
""
],
[
"Sud",
"James",
""
]
] |
2506.03442 | William Coon | William G. Coon, Preston Peranich, Griffin Milsap | StARS DCM: A Sleep Stage-Decoding Forehead EEG Patch for Real-time Modulation of Sleep Physiology | null | null | null | null | eess.SP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The System to Augment Restorative Sleep (StARS) is a modular hardware/software platform designed for real-time sleep monitoring and intervention. Utilizing the compact DCM biosignal device, StARS captures electrophysiological signals (EEG, EMG, EOG) and synchronizes sensor data using the ezmsg real-time software framework. StARS supports interventions such as closed-loop auditory stimulation and dynamic thermal modulation guided by sleep-stage decoding via advanced neural network models and transfer learning. Configurable with a lightweight EEG forehead patch or wearable sensors like smart rings, StARS offers flexible, low-burden solutions for EEG, BCI, and sleep-enhancement research and applications. The open-source DCM patch further enables customizable EEG device development. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:54:12 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Coon",
"William G.",
""
],
[
"Peranich",
"Preston",
""
],
[
"Milsap",
"Griffin",
""
]
] |
2506.03443 | Ali Salloum | Ali Salloum, Dorian Quelle, Letizia Iannucci, Alexandre Bovet, Mikko Kivel\"a | Politics and polarization on Bluesky | null | null | null | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | Online political discourse is increasingly shaped not by a few dominant platforms but by a fragmented ecosystem of social media spaces, each with its own user base, target audience, and algorithmic mediation of discussion. Such fragmentation may fundamentally change how polarization manifests online. In this study, we investigate the characteristics of political discourse and polarization on the emerging social media site Bluesky. We collect all activity on the platform between December 2024 and May 2025 to map out the platform's political topic landscape and detect distinct polarization patterns. Our comprehensive data collection allows us to employ a data-driven methodology for identifying political themes, classifying user stances, and measuring both structural and content-based polarization across key topics raised in English-language discussions. Our analysis reveals that approximately 13% of Bluesky posts engage with political content, with prominent topics including international conflicts, U.S. politics, and socio-technological debates. We find high levels of structural polarization across several salient political topics. However, the most polarized topics are also highly imbalanced in the numbers of users on opposing sides, with the smaller group consisting of only 1-2% of the users. While discussions in Bluesky echo familiar political narratives and polarization trends, the platform exhibits a more politically homogeneous user base than was typical prior to the current wave of platform fragmentation. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:54:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Salloum",
"Ali",
""
],
[
"Quelle",
"Dorian",
""
],
[
"Iannucci",
"Letizia",
""
],
[
"Bovet",
"Alexandre",
""
],
[
"Kivelä",
"Mikko",
""
]
] |
2506.03444 | Yue Gong | Yue Gong, Raul Castro Fernandez | Exploiting LLMs for Automatic Hypothesis Assessment via a Logit-Based Calibrated Prior | Under Review | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | As hypothesis generation becomes increasingly automated, a new bottleneck has emerged: hypothesis assessment. Modern systems can surface thousands of statistical relationships-correlations, trends, causal links-but offer little guidance on which ones are novel, non-trivial, or worthy of expert attention. In this work, we study the complementary problem to hypothesis generation: automatic hypothesis assessment. Specifically, we ask: given a large set of statistical relationships, can we automatically assess which ones are novel and worth further exploration? We focus on correlations as they are a common entry point in exploratory data analysis that often serve as the basis for forming deeper scientific or causal hypotheses.
To support automatic assessment, we propose to leverage the vast knowledge encoded in LLMs' weights to derive a prior distribution over the correlation value of a variable pair. If an LLM's prior expects the correlation value observed, then such correlation is not surprising, and vice versa. We propose the Logit-based Calibrated Prior, an LLM-elicited correlation prior that transforms the model's raw output logits into a calibrated, continuous predictive distribution over correlation values. We evaluate the prior on a benchmark of 2,096 real-world variable pairs and it achieves a sign accuracy of 78.8%, a mean absolute error of 0.26, and 95% credible interval coverage of 89.2% in predicting Pearson correlation coefficient. It also outperforms a fine-tuned RoBERTa classifier in binary correlation prediction and achieves higher precision@K in hypothesis ranking. We further show that the prior generalizes to correlations not seen during LLM pretraining, reflecting context-sensitive reasoning rather than memorization. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 22:54:59 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gong",
"Yue",
""
],
[
"Fernandez",
"Raul Castro",
""
]
] |
2506.03445 | Mohamed Cherifi | Mohamed Cherifi, Xujia Zhu, Mohammed Nabil El Korso and Ammar Mesloub | Maximum Likelihood for Logistic Regression Model with Incomplete and Hybrid-Type Covariates | 20 pages, 4 figures. To appear in IEEE Signal Processing Letters | null | null | null | stat.ME eess.SP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Logistic regression is a fundamental and widely used statistical method for modeling binary outcomes based on covariates. However, the presence of missing data, particularly in settings involving hybrid covariates (a mix of discrete and continuous variables), poses significant challenges. In this paper, we propose a novel Expectation-Maximization based algorithm tailored for parameter estimation in logistic regression models with missing hybrid covariates. The proposed method is specifically designed to handle these complexities, delivering efficient parameter estimates. Through comprehensive simulations and real-world application, we demonstrate that our approach consistently outperforms traditional methods, achieving superior accuracy and reliability. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:11:31 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cherifi",
"Mohamed",
""
],
[
"Zhu",
"Xujia",
""
],
[
"Korso",
"Mohammed Nabil El",
""
],
[
"Mesloub",
"Ammar",
""
]
] |
2506.03446 | John McHugh | John Revere McHugh | Isotypic blocks of finite groups algebras that are not $p$-permutation equivalent | null | null | null | null | math.RT math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that Kessar's isotypy between Galois conjugate blocks of finite group algebras does not always lift to a $p$-permutation equivalence. We also provide examples of Galois conjugate blocks which are isotypic but not $p$-permutation equivalent. These results help to clarify the distinction between a $p$-permutation equivalence and an isotypy, and may be useful in determining necessary and sufficient conditions for when an isotypy lifts to a $p$-permutation equivalence. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:15:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"McHugh",
"John Revere",
""
]
] |
2506.03447 | Marcos Gabriel Alpino | Marcos G. Alpino, Tiago Debarba, Reinaldo O. Vianna, Andr\'e T. Ces\'ario | A Complexity-Based Approach to Quantum Observable Equilibration | Comments: 11 pages, 8 figures. Comments are welcome | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the role of a statistical complexity measure to assign equilibration in isolated quantum systems. While unitary dynamics preserve global purity, expectation values of observables often exhibit equilibration-like behavior, raising the question of whether complexity can track this process. In addition to examining observable equilibration, we extend our analysis to study how the complexity of the quantum states evolves, providing insight into the transition from initial coherence to equilibrium. We define a classical statistical complexity measure based on observable entropy and deviation from equilibrium, which captures the dynamical progression towards equilibration and effectively distinguishes between complex and non-complex trajectories. In particular, our measure is sensitive to non-complex dynamics, such as the quasi-periodic behavior exhibited by low effective dimension initial states, where the systems explore a limited region of the Hilbert space as they oscillate in an informational coherence-preserving manner. These findings are supported by numerical simulations of an Ising-like non-integrable Hamiltonian spin-chain model. Our work provides new insight into the emergence of equilibrium behavior from unitary dynamics and advances complexity as a meaningful tool in the study of the emergence of classicality in microscopic systems. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:16:35 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Alpino",
"Marcos G.",
""
],
[
"Debarba",
"Tiago",
""
],
[
"Vianna",
"Reinaldo O.",
""
],
[
"Cesário",
"André T.",
""
]
] |
2506.03448 | Maitreya Patel | Bimsara Pathiraja, Maitreya Patel, Shivam Singh, Yezhou Yang, Chitta Baral | RefEdit: A Benchmark and Method for Improving Instruction-based Image Editing Model on Referring Expressions | Project page: \url{http://refedit.vercel.app} | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Despite recent advances in inversion and instruction-based image editing, existing approaches primarily excel at editing single, prominent objects but significantly struggle when applied to complex scenes containing multiple entities. To quantify this gap, we first introduce RefEdit-Bench, a rigorous real-world benchmark rooted in RefCOCO, where even baselines trained on millions of samples perform poorly. To overcome this limitation, we introduce RefEdit -- an instruction-based editing model trained on our scalable synthetic data generation pipeline. Our RefEdit, trained on only 20,000 editing triplets, outperforms the Flux/SD3 model-based baselines trained on millions of data. Extensive evaluations across various benchmarks demonstrate that our model not only excels in referring expression tasks but also enhances performance on traditional benchmarks, achieving state-of-the-art results comparable to closed-source methods. We release data \& checkpoint for reproducibility. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:20:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pathiraja",
"Bimsara",
""
],
[
"Patel",
"Maitreya",
""
],
[
"Singh",
"Shivam",
""
],
[
"Yang",
"Yezhou",
""
],
[
"Baral",
"Chitta",
""
]
] |
2506.03449 | John Smutny | John W. Smutny | The effects of using created synthetic images in computer vision training | Nine pages long. Main content in pages one through eight. References start at page nine | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | This paper investigates how rendering engines, like Unreal Engine 4 (UE), can be used to create synthetic images to supplement datasets for deep computer vision (CV) models in image abundant and image limited use cases. Using rendered synthetic images from UE can provide developers and businesses with a method of accessing nearly unlimited, reproducible, agile, and cheap training sets for their customers and applications without the threat of poisoned images from the internet or the cost of collecting them. The validity of these generated images are examined by testing the change in model test accuracy in two different sized CV models across two binary classification cases (Cat vs Dog and Weld Defect Detection). In addition, this paper provides an implementation of how to measure the quality of synthetic images by using pre-trained CV models as auditors. Results imply that for large (VGG16) and small (MobileNetV3-small) parameter deep CV models, adding >60% additional synthetic images to a real image dataset during model training can narrow the test-training accuracy gap to ~1-2% without a conclusive effect on test accuracy compared to using real world images alone. Likewise, adding <10% additional real training images to synthetic only training sets decreased the classification error rate in half, then decreasing further when adding more real training images. For these cases tested, using synthetic images from rendering engines allow researchers to only use 10% of their real images during training, compared to the traditional 50-70%. This research serves as an example of how to create synthetic images, guidelines on how to use the images, potential restrictions and possible performance improvements for data-scarce projects. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:20:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Smutny",
"John W.",
""
]
] |
2506.03450 | Prithvish Vijaykumar Nembhani | Prithvish V Nembhani, Oliver Rhodes, Guangzhi Tang, Alexandra F Dobrita, Yingfu Xu, Kanishkan Vadivel, Kevin Shidqi, Paul Detterer Mario Konijnenburg, Gert-Jan van Schaik, Manolis Sifalakis, Zaid Al-Ars, Amirreza Yousefzadeh | SENMAP: Multi-objective data-flow mapping and synthesis for hybrid scalable neuromorphic systems | IJCNN conference, italy, 2025, accepted, 30 June - 5 July | null | null | null | cs.NE | http://creativecommons.org/licenses/by/4.0/ | This paper introduces SENMap, a mapping and synthesis tool for scalable, energy-efficient neuromorphic computing architecture frameworks. SENECA is a flexible architectural design optimized for executing edge AI SNN/ANN inference applications efficiently. To speed up the silicon tape-out and chip design for SENECA, an accurate emulator, SENSIM, was designed. While SENSIM supports direct mapping of SNNs on neuromorphic architectures, as the SNN and ANNs grow in size, achieving optimal mapping for objectives like energy, throughput, area, and accuracy becomes challenging. This paper introduces SENMap, flexible mapping software for efficiently mapping large SNN and ANN applications onto adaptable architectures. SENMap considers architectural, pretrained SNN and ANN realistic examples, and event rate-based parameters and is open-sourced along with SENSIM to aid flexible neuromorphic chip design before fabrication. Experimental results show SENMap enables 40 percent energy improvements for a baseline SENSIM operating in timestep asynchronous mode of operation. SENMap is designed in such a way that it facilitates mapping large spiking neural networks for future modifications as well. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:21:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nembhani",
"Prithvish V",
""
],
[
"Rhodes",
"Oliver",
""
],
[
"Tang",
"Guangzhi",
""
],
[
"Dobrita",
"Alexandra F",
""
],
[
"Xu",
"Yingfu",
""
],
[
"Vadivel",
"Kanishkan",
""
],
[
"Shidqi",
"Kevin",
""
],
[
"Konijnenburg",
"Paul Detterer Mario",
""
],
[
"van Schaik",
"Gert-Jan",
""
],
[
"Sifalakis",
"Manolis",
""
],
[
"Al-Ars",
"Zaid",
""
],
[
"Yousefzadeh",
"Amirreza",
""
]
] |
2506.03451 | Kyriakos Destounis Dr. | H. S. Vieira and Kyriakos Destounis | Eye of the vortex: bound spectra in tunable horizonless rotational analogs | 12 pages, 3 figures | null | null | null | gr-qc astro-ph.GA hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | Analog gravity experiments are making remarkable strides in unveiling both the classical and quantum nature of black holes. By harnessing diverse states of matter, contemporary tabletop setups now replicate strong-field phenomena typically confined to the enigmatic regions surrounding black holes. Through these modern gravity simulators, physical processes once considered elusive may finally be brought into experimental reach. In this work, we investigate the bound spectrum of massless scalar excitations propagating within the effective geometry of a rotating acoustic metric. Specifically, we utilize an analog vortex endowed with a tunable parameter that emulates the spacetime of a rotating gravitational background. This model accommodates both the presence of a sonic horizon--characteristic of an acoustic black hole--for non-zero tuning parameters, and its absence when the parameter vanishes, yielding a horizonless, purely rotational vortex flow devoid of radial inflow. We focus on the latter case, where the vortex flow is purely rotational, and compute the spectral properties of the analog system. The resulting bound-state spectrum is found to be qualitatively consistent with that observed in recent experimental realizations of superfluid Helium giant quantum vortices featuring solid or hollow cores. This correspondence suggests that the analog spacetime geometry used here holds significant potential to replicate the phenomenology of cutting-edge laboratory experiments. In doing so, it offers new insight into the intricate landscape of analog black hole spectroscopy and, potentially, the physical topography of bounded, rotating astrophysical environments around black holes. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:27:11 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Vieira",
"H. S.",
""
],
[
"Destounis",
"Kyriakos",
""
]
] |
2506.03452 | Timothy Mahoney | Timothy B Mahoney, JingYang Liu, Huakun Xin, David B Grayden, Sam E John | Minimally Invasive Brain Computer Interfaces: Evaluating the Impact of Tissue Layers on Signal Quality of Sub-Scalp EEG | 20 pages, 8 figures, 2 tables | null | null | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | Individuals with severe physical disabilities often experience diminished quality of life stemming from limited ability to engage with their surroundings. Brain-Computer Interface (BCI) technology aims to bridge this gap by enabling direct technology interaction. However, current BCI systems require invasive procedures, such as craniotomy or implantation of electrodes through blood vessels, posing significant risks to patients. Sub-scalp electroencephalography (EEG) offers a lower risk alternative. This study investigates the signal quality of sub-scalp EEG recordings from various depths in a sheep model, and compares results with other methods: ECoG and endovascular arrays. A computational model was also constructed to investigate the factors underlying variations in electrode performance. We demonstrate that peg electrodes placed within the sub-scalp space can achieve visual evoked potential signal-to-noise ratios (SNRs) approaching that of ECoG. Endovascular arrays exhibited SNR comparable to electrodes positioned on the periosteum. Furthermore, sub-scalp recordings captured high gamma neural activity, with maximum bandwidth ranging from 120 Hz to 180 Hz depending on electrode depth. These findings support the use of sub-scalp EEG for BCI applications, and provide valuable insights for future sub-scalp electrode design. This data lays the groundwork for human trials, ultimately paving the way for chronic, in-home BCIs that empower individuals with physical disabilities. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:30:21 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mahoney",
"Timothy B",
""
],
[
"Liu",
"JingYang",
""
],
[
"Xin",
"Huakun",
""
],
[
"Grayden",
"David B",
""
],
[
"John",
"Sam E",
""
]
] |
2506.03453 | Iman Marvian | Plato Deliyannis and Iman Marvian | Permutationally-Invariant N-body gates via Tavis-Cummings Hamiltonian | Preliminary version; 18 pages + 33 pages of Appendices + 9 Figures; Comments Welcome! | null | null | null | quant-ph math-ph math.MP nucl-th physics.atom-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Widely used in atomic and superconducting qubit systems, the Jaynes-Cummings (JC) Hamiltonian is a simple, yet powerful model for a two-level system interacting with a quantum harmonic oscillator. In this paper, we focus on a system of n qubits, identically coupled to a single oscillator via JC interaction, also known as the Tavis-Cummings (TC) Hamiltonian. We show that all permutationally-invariant unitaries on an arbitrary number of qubits can be realized using this permutationally-invariant Hamiltonian, which couples the qubits to an oscillator initialized in its vacuum state, together with global uniform x and z fields on all qubits. This includes useful gates, such as controlled-Z gate with an arbitrary number of control qubits. As a corollary, we find that all permutationally invariant states -- including useful entangled states such as GHZ and Dicke states -- can be prepared using this interaction and global fields. We also characterize unitaries that can be realized on the joint Hilbert space of the qubits and oscillator with the TC interaction and global z field, and develop new methods for preparing the state of the oscillator in an arbitrary initial state. We present various examples of explicit circuits for the case of n=2 qubits. In particular, we develop new methods for implementing controlled-Z, SWAP, iSWAP, and $\sqrt{i\text{SWAP}}$ gates using only the TC interaction and a global z field. Our work also reveals an accidental symmetry in the TC Hamiltonian and shows that it can be explained using Schwinger's oscillator model of angular momentum. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:33:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Deliyannis",
"Plato",
""
],
[
"Marvian",
"Iman",
""
]
] |
2506.03454 | Muratkhan Abdirash | Muratkhan Abdirash, Xiaofan Cui | Nonlinear Optimal Control of DC Microgrids with Safety and Stability Guarantees | To appear in the proceedings of American Control Conference (ACC) 2025 | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | A DC microgrid is a promising alternative to the traditional AC power grid, since it can efficiently integrate distributed and renewable energy resources. However, as an emerging framework, it lacks the rigorous theoretical guarantees of its AC counterpart. In particular, safe stabilization of the DC microgrid has been a non-trivial task in power electronics. To address that, we take a control theoretic perspective in designing the feedback controller with provable guarantees. We present a systematic way to construct Control Lyapunov Functions (CLF) to stabilize the microgrid, and, independently, Control Barrier Functions (CBF) to enforce its safe operation at all times. The safety-critical controller (SCC) proposed in this work integrates the two control objectives, with safety prioritized, into a quadratic program (QP) as linear constraints, which allows for its online deployment using off-the-shelf convex optimization solvers. The SCC is compared against a robust version of the conventional droop control through numerical experiments whose results indicate the SCC outperforms the droop controller in guaranteeing safety and retaining stability at the same time. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:37:28 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Abdirash",
"Muratkhan",
""
],
[
"Cui",
"Xiaofan",
""
]
] |
2506.03455 | Ariel Norambuena | Hachisko Tapia-Maureira, Bing He, Massimiliano Di Ventra, and Ariel Norambuena | Memory effects in pulsed optomechanical systems | 12 pages, 6 figures, 1 table | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Memory, understood as time non-locality, is a fundamental property of any physical system, whether classical or quantum, and has important applications in a wide variety of technologies. In the context of quantum technologies, systems with memory can be used in quantum information, communication, and sensing. Here, we demonstrate that cavity optomechanical systems driven by a pulsed laser can operate as programmable quantum memory elements. By engineering the adiabatic and non-adiabatic pulses, particularly the Gaussian and sinusoidal, we induce and control diverse memory phenomena such as dynamical hysteresis, quantized phononic transitions, and distinct energy-storing responses. Within a mean-field approach, we derive the analytical and numerical criteria under which the photonic and phononic observables manifest the memory effects in strongly driven regimes. The memory effects are quantified through a dimensionless geometric form factor, which provides a versatile metric to characterize the memory efficiency. Our protocol is readily compatible with the current optomechanical platforms, highlighting the new possibilities for advanced memory functionalities in quantum technologies. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:39:02 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Tapia-Maureira",
"Hachisko",
""
],
[
"He",
"Bing",
""
],
[
"Di Ventra",
"Massimiliano",
""
],
[
"Norambuena",
"Ariel",
""
]
] |
2506.03456 | Mingkang Xia | Mingkang Xia, Crist\'obal Lled\'o, Matthew Capocci, Jacob Repicky, Benjamin D'Anjou, Ian Mondragon-Shem, Ryan Kaufman, Jens Koch, Alexandre Blais, Michael Hatridge | Exceeding the Parametric Drive Strength Threshold in Nonlinear Circuits | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Superconducting quantum circuits rely on strong drives to implement fast gates, high-fidelity readout, and state stabilization. However, these drives can induce uncontrolled excitations, so-called "ionization", that compromise the fidelity of these operations. While now well-characterized in the context of qubit readout, it remains unclear how general this limitation is across the more general setting of parametric control. Here, we demonstrate that a nonlinear coupler, exemplified by a transmon, undergoes ionization under strong parametric driving, leading to a breakdown of coherent control and thereby limiting the accessible gate speeds. Through experiments and numerical simulations, we associate this behavior with the emergence of drive-induced chaotic dynamics, which we characterize quantitatively using the instantaneous Floquet spectrum. Our results reveal that the Floquet spectrum provides a unifying framework for understanding strong-drive limitations across a wide range of operations on superconducting quantum circuits. This insight establishes fundamental constraints on parametric control and offers design principles for mitigating drive-induced decoherence in next-generation quantum processors. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:39:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xia",
"Mingkang",
""
],
[
"Lledó",
"Cristóbal",
""
],
[
"Capocci",
"Matthew",
""
],
[
"Repicky",
"Jacob",
""
],
[
"D'Anjou",
"Benjamin",
""
],
[
"Mondragon-Shem",
"Ian",
""
],
[
"Kaufman",
"Ryan",
""
],
[
"Koch",
"Jens",
""
],
[
"Blais",
"Alexandre",
""
],
[
"Hatridge",
"Michael",
""
]
] |
2506.03457 | Lei Bill Wang | Lei Bill Wang, Sooa Ahn | Disentangling Barriers to Welfare Program Participation with Semiparametric and Mixed Effect Approaches | null | null | null | null | econ.GN q-fin.EC | http://creativecommons.org/licenses/by/4.0/ | This paper examines why eligible households do not participate in welfare programs. Under the assumption that there exist some observed fully attentive groups, we model take-up as a two-stage process: attention followed by choice. We do so with two novel approaches. Drawing inspiration from the demand estimation for stochastically attentive consumers literature, Approach I is semiparametric with a nonparametric attention function and a parametric choice function. It uses fully attentive households to identify choice utility parameters and then uses the entire population to identify the attention probabilities. By augmenting Approach I with a random effect that simultaneously affects the attention and choice stages, Approach II allows household-level unobserved heterogeneity and dependence between attention and choice even after conditioning on observed covariates.
Applied to NLSY panel data for WIC participation, both approaches consistently point to two empirical findings with regard to heterogeneous policy targeting. (1) As an infant ages towards 12 months and beyond, attention probability drops dramatically while choice probability steadily decreases. Finding (1) suggests that exit-prevention is the key for increasing the take-up rate because once a household exits the program when the infant ages close to 12 months old, it is unlikely to rejoin due to low attention. A value-increasing solution is predicted to be effective in promoting take-up by reducing exit probability. In contrast, an attention-raising solution is predicted to be ineffective. (2) Higher educated households are less attentive but more likely to enroll if attentive. Finding (2) suggests that running informational campaigns with parenting student groups at higher education institutions could be an effective strategy for boosting take-up. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:40:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Lei Bill",
""
],
[
"Ahn",
"Sooa",
""
]
] |
2506.03458 | Zahra Bokaei | Zahra Bokaei, Walid Magdy, Bonnie Webber | Culture Matters in Toxic Language Detection in Persian | Accepted to ACL 2025 (Main Track) | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Toxic language detection is crucial for creating safer online environments and limiting the spread of harmful content. While toxic language detection has been under-explored in Persian, the current work compares different methods for this task, including fine-tuning, data enrichment, zero-shot and few-shot learning, and cross-lingual transfer learning. What is especially compelling is the impact of cultural context on transfer learning for this task: We show that the language of a country with cultural similarities to Persian yields better results in transfer learning. Conversely, the improvement is lower when the language comes from a culturally distinct country. Warning: This paper contains examples of toxic language that may disturb some readers. These examples are included for the purpose of research on toxic detection. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:48:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bokaei",
"Zahra",
""
],
[
"Magdy",
"Walid",
""
],
[
"Webber",
"Bonnie",
""
]
] |
2506.03459 | Garud Snoep | G. Snoep, C. Bourdelle, J. Citrin, A. Ho, M. J. Pueschel, P. Vincenzi, E. R. Solano, M. Sertoli, E. Delabie and Jet Contributors | Characterization of reduced-order turbulence models in the L-mode pedestal-forming region in JET | null | null | null | null | physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Linear instability characterization of seven JET discharges just prior to the L-H transition is performed at $\rho_{\text{tor}} \in [0.85,0.9,0.95]$ with the gyrokinetic GENE code. The discharges cover both the low- and high-density branches of the L-H transition at two different triangularities. Sensitivities to driving gradients, normalized electron collisonality $\nu_e^*$, hydrogen isotope mass, magnetic geometry and finite-$\beta$ effects are all characterized. At $\rho_{\text{tor}}=0.85$ and $0.9$, trapped-electron modes (TEMs) propagating in both the electron- or ion-drift direction are observed at the lowest densities. At higher density ion-temperature-gradient (ITG) modes are dominant, some of which exhibit trapped-ion drive and unconventional ballooning structures. At $\rho_{\text{tor}}=0.95$, the low-density cases are similar to inner radii, while at higher densities subdominant modes are destabilized by higher collisionalities. The electron collisonality $\nu_e^*$ is scanned around the experimental values at all three radii and for the seven discharges studied. The experimental collisionality range corresponds to a region of minimum linear drive between ITG-TEM mode branches at lower collisionalities and resistive mode branches at higher collisionalities. Moreover, the quasilinear particle flux is directed inward only in the collisionality domain where the linear drive is minimized at $\rho_{\text{tor}}=0.85$ for all densities and $0.9$ only for the highest densities. Model fidelity reduction is performed on the GENE simulations to evaluate the impact of various assumptions and simplifications made by the state-of-the-art quasilinear models QuaLiKiz and TGLF. QuaLiKiz is found to be inadequate beyond $\rho_{\text{tor}}=0.85$, while TGLF-SAT2 agrees well with linear spectra and the quasilinear heat fluxes from GENE up to and including $\rho_{\text{tor}}=0.9$. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:55:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Snoep",
"G.",
""
],
[
"Bourdelle",
"C.",
""
],
[
"Citrin",
"J.",
""
],
[
"Ho",
"A.",
""
],
[
"Pueschel",
"M. J.",
""
],
[
"Vincenzi",
"P.",
""
],
[
"Solano",
"E. R.",
""
],
[
"Sertoli",
"M.",
""
],
[
"Delabie",
"E.",
""
],
[
"Contributors",
"Jet",
""
]
] |
2506.03461 | Dequan Jin | Nan Xiang, Lifeng Xing, Dequan Jin | RoNFA: Robust Neural Field-based Approach for Few-Shot Image Classification with Noisy Labels | 7 pages, 1 figure | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In few-shot learning (FSL), the labeled samples are scarce. Thus, label errors can significantly reduce classification accuracy. Since label errors are inevitable in realistic learning tasks, improving the robustness of the model in the presence of label errors is critical. This paper proposes a new robust neural field-based image approach (RoNFA) for few-shot image classification with noisy labels. RoNFA consists of two neural fields for feature and category representation. They correspond to the feature space and category set. Each neuron in the field for category representation (FCR) has a receptive field (RF) on the field for feature representation (FFR) centered at the representative neuron for its category generated by soft clustering. In the prediction stage, the range of these receptive fields adapts according to the neuronal activation in FCR to ensure prediction accuracy. These learning strategies provide the proposed model with excellent few-shot learning capability and strong robustness against label noises. The experimental results on real-world FSL datasets with three different types of label noise demonstrate that the proposed method significantly outperforms state-of-the-art FSL methods. Its accuracy obtained in the presence of noisy labels even surpasses the results obtained by state-of-the-art FSL methods trained on clean support sets, indicating its strong robustness against noisy labels. | [
{
"version": "v1",
"created": "Tue, 03 Jun 2025 23:57:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xiang",
"Nan",
""
],
[
"Xing",
"Lifeng",
""
],
[
"Jin",
"Dequan",
""
]
] |
2506.03462 | Yeonjoo Park | Yeonjoo Park and Aiguo Han | Robust domain selection for functional data via interval-wise testing and effect size mapping | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Among inferential problems in functional data analysis, domain selection is one of the practical interests aiming to identify sub-interval(s) of the domain where desired functional features are displayed. Motivated by applications in quantitative ultrasound signal analysis, we propose the robust domain selection method, particularly aiming to discover a subset of the domain presenting distinct behaviors on location parameters among different groups. By extending the interval testing approach, we propose to take into account multiple aspects of functional features simultaneously to detect the practically interpretable domain. To further handle potential outliers and missing segments on collected functional trajectories, we perform interval testing with a test statistic based on functional M-estimators for the inference. In addition, we introduce the effect size heatmap by calculating robustified effect sizes from the lowest to the largest scales over the domain to reflect dynamic functional behaviors among groups so that clinicians get a comprehensive understanding and select practically meaningful sub-interval(s). The performance of the proposed method is demonstrated through simulation studies and an application to motivating quantitative ultrasound measurements. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:01:16 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Park",
"Yeonjoo",
""
],
[
"Han",
"Aiguo",
""
]
] |
2506.03463 | Zhigang Wang | Jia Fan, Chenxi Wan, Rui Liu, Zhen Gong, Hongbo Jing, Baiqiang Liu, Siyang Liu, and Zhigang Wang | Superatomic hydrogen: achieving effective aggregation of hydrogen atoms at pressures lower than that of metallic hydrogen | 7 pages, 3 figures | null | null | null | physics.comp-ph cond-mat.str-el physics.atom-ph physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Metal hydrogen exhibiting electron delocalization properties has been recognized as an important prospect for achieving controlled nuclear fusion, but the extreme pressure conditions required exceeding hundreds of GPa remain a daunting challenge. Here, we propose a model of superatomic hydrogen, aiming to reduce the pressure conditions required for the effective aggregation of elemental hydrogen atoms. High-precision ab initio calculations indicate that the pressure required to compress the H13 system with one central atom and 12 surrounding atoms into a superatomic state is approximately two orders of magnitude lower than that of metallic hydrogen. Atomic-level analyses reveal that in the superatomic state of compressed H13, the central H atom donates its electron, and all electrons are delocalized on the superatomic molecular orbitals, which conforms to properties of metallic hydrogen. Our discovery in principle opens up the prospect of superatomic hydrogen in areas such as nuclear fusion. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:23:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fan",
"Jia",
""
],
[
"Wan",
"Chenxi",
""
],
[
"Liu",
"Rui",
""
],
[
"Gong",
"Zhen",
""
],
[
"Jing",
"Hongbo",
""
],
[
"Liu",
"Baiqiang",
""
],
[
"Liu",
"Siyang",
""
],
[
"Wang",
"Zhigang",
""
]
] |
2506.03464 | Weiqiang Zheng | Yang Cai, Haipeng Luo, Chen-Yu Wei, Weiqiang Zheng | From Average-Iterate to Last-Iterate Convergence in Games: A Reduction and Its Applications | 21 pages | null | null | null | cs.GT cs.LG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The convergence of online learning algorithms in games under self-play is a fundamental question in game theory and machine learning. Among various notions of convergence, last-iterate convergence is particularly desirable, as it reflects the actual decisions made by the learners and captures the day-to-day behavior of the learning dynamics. While many algorithms are known to converge in the average-iterate, achieving last-iterate convergence typically requires considerably more effort in both the design and the analysis of the algorithm. Somewhat surprisingly, we show in this paper that for a large family of games, there exists a simple black-box reduction that transforms the average iterates of an uncoupled learning dynamics into the last iterates of a new uncoupled learning dynamics, thus also providing a reduction from last-iterate convergence to average-iterate convergence. Our reduction applies to games where each player's utility is linear in both their own strategy and the joint strategy of all opponents. This family includes two-player bimatrix games and generalizations such as multi-player polymatrix games. By applying our reduction to the Optimistic Multiplicative Weights Update algorithm, we obtain new state-of-the-art last-iterate convergence rates for uncoupled learning dynamics in two-player zero-sum normal-form games: (1) an $O(\frac{\log d}{T})$ last-iterate convergence rate under gradient feedback, representing an exponential improvement in the dependence on the dimension $d$ (i.e., the maximum number of actions available to either player); and (2) an $\widetilde{O}(d^{\frac{1}{5}} T^{-\frac{1}{5}})$ last-iterate convergence rate under bandit feedback, improving upon the previous best rates of $\widetilde{O}(\sqrt{d} T^{-\frac{1}{8}})$ and $\widetilde{O}(\sqrt{d} T^{-\frac{1}{6}})$. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:24:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cai",
"Yang",
""
],
[
"Luo",
"Haipeng",
""
],
[
"Wei",
"Chen-Yu",
""
],
[
"Zheng",
"Weiqiang",
""
]
] |
2506.03465 | Ashay Burungale | Ashay A. Burungale, Ye Tian | A rank zero $p$-converse to a theorem of Gross--Zagier, Kolyvagin and Rubin | This is a preprint from December 2019, submitted then and revised as per a referee report. See also https://ems.press/books/standalone/275/5456 | null | null | null | math.NT | http://creativecommons.org/licenses/by/4.0/ | Let $E$ be a CM elliptic curve defined over $\mathbb{Q}$ and $p$ a prime. We show that $${\mathrm corank}_{\mathbb{Z}_{p}} {\mathrm Sel}_{p^{\infty}}(E_{/\mathbb{Q}})=0 \implies {\mathrm ord}_{s=1}L(s,E_{/\mathbb{Q}})=0 $$ for the $p^{\infty}$-Selmer group ${\mathrm Sel}_{p^{\infty}}(E_{/\mathbb{Q}})$ and the complex $L$-function $L(s,E_{/\mathbb{Q}})$. Along with Smith's work on the distribution of $2^\infty$-Selmer groups, this leads to the first instance of the even parity Goldfeld conjecture: For $50\%$ of the positive square-free integers $n$, we have $ {\mathrm ord}_{s=1}L(s,E^{(n)}_{/\mathbb{Q}})=0, $ where $E^{(n)}: ny^{2}=x^{3}-x $ is a quadratic twist of the congruent number elliptic curve $E: y^{2}=x^{3}-x$. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:29:36 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Burungale",
"Ashay A.",
""
],
[
"Tian",
"Ye",
""
]
] |
2506.03466 | Hengrui Luo | James Demmel, Hengrui Luo, Ryan Schneider, Yifu Wang | Minimizing the Arithmetic and Communication Complexity of Jacobi's Method for Eigenvalues and Singular Values | 33 pages, 5 figures, 3 tables | null | null | null | math.NA cs.CC cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we analyze several versions of Jacobi's method for the symmetric eigenvalue problem. Our goal throughout is to reduce the asymptotic cost of the algorithm as much as possible, as measured by the number of arithmetic operations performed and associated (sequential or parallel) communication, i.e., the amount of data moved between slow and fast memory or between processors in a network. In producing rigorous complexity bounds, we allow our algorithms to be built on both classic $O(n^3)$ matrix multiplication and fast, Strassen-like $O(n^{\omega_0})$ alternatives. In the classical setting, we show that a blocked implementation of Jacobi's method attains the communication lower bound for $O(n^3)$ matrix multiplication (and is therefore expected to be communication optimal among $O(n^3)$ methods). In the fast setting, we demonstrate that a recursive version of blocked Jacobi can go even further, reaching essentially optimal complexity in both measures. We also discuss Jacobi-based SVD algorithms and a parallel version of block Jacobi, showing that analogous complexity bounds apply. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:38:56 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Demmel",
"James",
""
],
[
"Luo",
"Hengrui",
""
],
[
"Schneider",
"Ryan",
""
],
[
"Wang",
"Yifu",
""
]
] |
2506.03467 | Hang Liu | Hang Liu, Anna Scaglione, and Sean Peisert | Differentially Private Distribution Release of Gaussian Mixture Models via KL-Divergence Minimization | This work has been submitted to the IEEE for possible publication | null | null | null | cs.IT cs.CR cs.LG eess.SP math.IT stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gaussian Mixture Models (GMMs) are widely used statistical models for representing multi-modal data distributions, with numerous applications in data mining, pattern recognition, data simulation, and machine learning. However, recent research has shown that releasing GMM parameters poses significant privacy risks, potentially exposing sensitive information about the underlying data. In this paper, we address the challenge of releasing GMM parameters while ensuring differential privacy (DP) guarantees. Specifically, we focus on the privacy protection of mixture weights, component means, and covariance matrices. We propose to use Kullback-Leibler (KL) divergence as a utility metric to assess the accuracy of the released GMM, as it captures the joint impact of noise perturbation on all the model parameters. To achieve privacy, we introduce a DP mechanism that adds carefully calibrated random perturbations to the GMM parameters. Through theoretical analysis, we quantify the effects of privacy budget allocation and perturbation statistics on the DP guarantee, and derive a tractable expression for evaluating KL divergence. We formulate and solve an optimization problem to minimize the KL divergence between the released and original models, subject to a given $(\epsilon, \delta)$-DP constraint. Extensive experiments on both synthetic and real-world datasets demonstrate that our approach achieves strong privacy guarantees while maintaining high utility. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:40:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Liu",
"Hang",
""
],
[
"Scaglione",
"Anna",
""
],
[
"Peisert",
"Sean",
""
]
] |
2506.03468 | Stanley Lazic | Stanley E. Lazic | Internal replication as a tool for evaluating reproducibility in preclinical experiments | null | null | null | null | stat.AP q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Reproducibility is central to the credibility of scientific findings, yet complete replication studies are costly and infrequent. However, many biological experiments contain internal replication, which is defined as repetition across batches, runs, days, litters, or sites that can be used to estimate reproducibility without requiring additional experiments. This internal replication is analogous to internal validation in prediction or machine learning models, but is often treated as a nuisance and removed by normalisation, missing an opportunity to assess the stability of results. Here, six types of internal replication are defined based on independence and timing. Using mice data from an experiment conducted at three independent sites, we demonstrate how to quantify and test for internal reproducibility. This approach provides a framework for quantifying reproducibility from existing data and reporting more robust statistical inferences in preclinical research. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:50:42 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Lazic",
"Stanley E.",
""
]
] |
2506.03469 | Tuan Le | Tuan Le, Risal Shefin, Debashis Gupta, Thai Le and Sarra Alqahtani | Verification-Guided Falsification for Safe RL via Explainable Abstraction and Risk-Aware Exploration | 8 pages, 7 figures, European Conference on Artificial Intelligence (ECAI) | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Ensuring the safety of reinforcement learning (RL) policies in high-stakes environments requires not only formal verification but also interpretability and targeted falsification. While model checking provides formal guarantees, its effectiveness is limited by abstraction quality and the completeness of the underlying trajectory dataset. We propose a hybrid framework that integrates (1) explainability, (2) model checking, and (3) risk-guided falsification to achieve both rigor and coverage. Our approach begins by constructing a human-interpretable abstraction of the RL policy using Comprehensible Abstract Policy Summarization (CAPS). This abstract graph, derived from offline trajectories, is both verifier-friendly, semantically meaningful, and can be used as input to Storm probabilistic model checker to verify satisfaction of temporal safety specifications. If the model checker identifies a violation, it will return an interpretable counterexample trace by which the policy fails the safety requirement. However, if no violation is detected, we cannot conclude satisfaction due to potential limitation in the abstraction and coverage of the offline dataset. In such cases, we estimate associated risk during model checking to guide a falsification strategy that prioritizes searching in high-risk states and regions underrepresented in the trajectory dataset. We further provide PAC-style guarantees on the likelihood of uncovering undetected violations. Finally, we incorporate a lightweight safety shield that switches to a fallback policy at runtime when such a risk exceeds a threshold, facilitating failure mitigation without retraining. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:54:01 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Le",
"Tuan",
""
],
[
"Shefin",
"Risal",
""
],
[
"Gupta",
"Debashis",
""
],
[
"Le",
"Thai",
""
],
[
"Alqahtani",
"Sarra",
""
]
] |
2506.03470 | Liam Hodgkinson | Liam Hodgkinson, Zhichao Wang, and Michael W. Mahoney | Models of Heavy-Tailed Mechanistic Universality | 40 pages, 4 figures | null | null | null | stat.ML cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent theoretical and empirical successes in deep learning, including the celebrated neural scaling laws, are punctuated by the observation that many objects of interest tend to exhibit some form of heavy-tailed or power law behavior. In particular, the prevalence of heavy-tailed spectral densities in Jacobians, Hessians, and weight matrices has led to the introduction of the concept of heavy-tailed mechanistic universality (HT-MU). Multiple lines of empirical evidence suggest a robust correlation between heavy-tailed metrics and model performance, indicating that HT-MU may be a fundamental aspect of deep learning efficacy. Here, we propose a general family of random matrix models -- the high-temperature Marchenko-Pastur (HTMP) ensemble -- to explore attributes that give rise to heavy-tailed behavior in trained neural networks. Under this model, spectral densities with power laws on (upper and lower) tails arise through a combination of three independent factors (complex correlation structures in the data; reduced temperatures during training; and reduced eigenvector entropy), appearing as an implicit bias in the model structure, and they can be controlled with an "eigenvalue repulsion" parameter. Implications of our model on other appearances of heavy tails, including neural scaling laws, optimizer trajectories, and the five-plus-one phases of neural network training, are discussed. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:55:01 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hodgkinson",
"Liam",
""
],
[
"Wang",
"Zhichao",
""
],
[
"Mahoney",
"Michael W.",
""
]
] |
2506.03471 | Dongwook Lee | Christopher DeGrendele and Dongwook Lee | GP-Recipe: Gaussian Process approximation to linear operations in numerical methods | null | null | null | null | physics.comp-ph cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | We introduce new Gaussian Process (GP) high-order approximations to linear operations that are frequently used in various numerical methods. Our method employs the kernel-based GP regression modeling, a non-parametric Bayesian approach to regression that operates on the probability distribution over all admissible functions that fit observed data. We begin in the first part with discrete data approximations to various linear operators applied to smooth data using the most popular squared exponential kernel function. In the second part, we discuss data interpolation across discontinuities with sharp gradients, for which we introduce a new GP kernel that fits discontinuous data without oscillations. The current study extends our previous GP work on polynomial-free shock-capturing methods in finite difference and finite volume methods to a suite of linear operator approximations on smooth data. The formulations introduced in this paper can be readily adopted in daily practices in numerical methods, including numerical approximations of finite differences, quadrature rules, interpolations, and reconstructions, which are most frequently used in numerical modeling in modern science and engineering applications. In the test problems, we demonstrate that the GP approximated solutions feature improved solution accuracy compared to the conventional finite-difference counterparts. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 00:59:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"DeGrendele",
"Christopher",
""
],
[
"Lee",
"Dongwook",
""
]
] |
2506.03472 | Mahesh Godavarti | Mahesh Godavarti | Directional Non-Commutative Monoidal Embeddings for MNIST | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We present an empirical validation of the directional non-commutative monoidal embedding framework recently introduced in prior work~\cite{Godavarti2025monoidal}. This framework defines learnable compositional embeddings using distinct non-commutative operators per dimension (axis) that satisfy an interchange law, generalizing classical one-dimensional transforms. Our primary goal is to verify that this framework can effectively model real data by applying it to a controlled, well-understood task: image classification on the MNIST dataset~\cite{lecun1998gradient}. A central hypothesis for why the proposed monoidal embedding works well is that it generalizes the Discrete Fourier Transform (DFT)~\cite{oppenheim1999discrete} by learning task-specific frequency components instead of using fixed basis frequencies. We test this hypothesis by comparing learned monoidal embeddings against fixed DFT-based embeddings on MNIST. The results show that as the embedding dimensionality decreases (e.g., from 32 to 8 to 2), the performance gap between the learned monoidal embeddings and fixed DFT-based embeddings on MNIST grows increasingly large. This comparison is used as an analytic tool to explain why the framework performs well: the learnable embeddings can capture the most discriminative spectral components for the task. Overall, our experiments confirm that directional non-commutative monoidal embeddings are highly effective for representing image data, offering a compact learned representation that retains high task performance. The code used in this work is available at https://github.com/mahesh-godavarti/directional_composition_mnist. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:07:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Godavarti",
"Mahesh",
""
]
] |
2506.03473 | Xinru Ying | Xinru Ying, Jiaqi Mo, Jingyang Lin, Canghong Jin, Fangfang Wang, Lina Wei | MamFusion: Multi-Mamba with Temporal Fusion for Partially Relevant Video Retrieval | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Partially Relevant Video Retrieval (PRVR) is a challenging task in the domain of multimedia retrieval. It is designed to identify and retrieve untrimmed videos that are partially relevant to the provided query. In this work, we investigate long-sequence video content understanding to address information redundancy issues. Leveraging the outstanding long-term state space modeling capability and linear scalability of the Mamba module, we introduce a multi-Mamba module with temporal fusion framework (MamFusion) tailored for PRVR task. This framework effectively captures the state-relatedness in long-term video content and seamlessly integrates it into text-video relevance understanding, thereby enhancing the retrieval process. Specifically, we introduce Temporal T-to-V Fusion and Temporal V-to-T Fusion to explicitly model temporal relationships between text queries and video moments, improving contextual awareness and retrieval accuracy. Extensive experiments conducted on large-scale datasets demonstrate that MamFusion achieves state-of-the-art performance in retrieval effectiveness. Code is available at the link: https://github.com/Vision-Multimodal-Lab-HZCU/MamFusion. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:08:03 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ying",
"Xinru",
""
],
[
"Mo",
"Jiaqi",
""
],
[
"Lin",
"Jingyang",
""
],
[
"Jin",
"Canghong",
""
],
[
"Wang",
"Fangfang",
""
],
[
"Wei",
"Lina",
""
]
] |
2506.03474 | Yifeng Xiao | Yifeng Xiao, Yurong Xu, Ning Yan, Masood Mortazavi, Pierluigi Nuzzo | CORE: Constraint-Aware One-Step Reinforcement Learning for Simulation-Guided Neural Network Accelerator Design | Preprint. 10 pages + appendix. Submitted to NeurIPS 2025 | null | null | null | cs.LG cs.AI cs.AR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Simulation-based design space exploration (DSE) aims to efficiently optimize high-dimensional structured designs under complex constraints and expensive evaluation costs. Existing approaches, including heuristic and multi-step reinforcement learning (RL) methods, struggle to balance sampling efficiency and constraint satisfaction due to sparse, delayed feedback, and large hybrid action spaces. In this paper, we introduce CORE, a constraint-aware, one-step RL method for simulationguided DSE. In CORE, the policy agent learns to sample design configurations by defining a structured distribution over them, incorporating dependencies via a scaling-graph-based decoder, and by reward shaping to penalize invalid designs based on the feedback obtained from simulation. CORE updates the policy using a surrogate objective that compares the rewards of designs within a sampled batch, without learning a value function. This critic-free formulation enables efficient learning by encouraging the selection of higher-reward designs. We instantiate CORE for hardware-mapping co-design of neural network accelerators, demonstrating that it significantly improves sample efficiency and achieves better accelerator configurations compared to state-of-the-art baselines. Our approach is general and applicable to a broad class of discrete-continuous constrained design problems. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:08:34 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xiao",
"Yifeng",
""
],
[
"Xu",
"Yurong",
""
],
[
"Yan",
"Ning",
""
],
[
"Mortazavi",
"Masood",
""
],
[
"Nuzzo",
"Pierluigi",
""
]
] |
2506.03475 | Zhijie Chen | Zhijie Chen | On the distribution of critical points of the Eisenstein series $E_6$ and monodromy interpretation | null | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In previous works joint with Lin, we proved that the Eisenstein series $E_4$ (resp. $E_2$) has at most one critical point in every fundamental domain $\gamma(F_0)$ of $\Gamma_{0}(2)$, where $\gamma(F_0)$ are translates of the basic fundamental domain $F_0$ via the M\"{o}bius transformation of $\gamma\in\Gamma_{0}(2)$. But the method can not work for the Eisenstein series $E_6$.
In this paper, we develop a new approach to show that $E_6'(\tau)$ has exactly either $1$ or $2$ zeros in every fundamental domain $\gamma(F_0)$ of $\Gamma_{0}(2)$. A criterion for $\gamma(F_0)$ containing exactly $2$ zeros is also given. Furthermore, by mapping all zeros of $E_6'(\tau)$ into $F_0$ via the M\"{o}bius transformations of $\Gamma_{0}(2)$ action, the images give rise to a dense subset of the union of three disjoint smooth curves in $F_0$. A monodromy interpretation of these curves from a complex linear ODE is also given. As a consequence, we give a complete description of the distribution of the zeros of $E_6'(\tau)$ in fundamental domains of $SL(2,\mathbb{Z})$. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:10:36 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Zhijie",
""
]
] |
2506.03476 | Chuyuan Li | Chuyuan Li, Raymond Li, Thalia S. Field, Giuseppe Carenini | Delta-KNN: Improving Demonstration Selection in In-Context Learning for Alzheimer's Disease Detection | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Alzheimer's Disease (AD) is a progressive neurodegenerative disorder that leads to dementia, and early intervention can greatly benefit from analyzing linguistic abnormalities. In this work, we explore the potential of Large Language Models (LLMs) as health assistants for AD diagnosis from patient-generated text using in-context learning (ICL), where tasks are defined through a few input-output examples. Empirical results reveal that conventional ICL methods, such as similarity-based selection, perform poorly for AD diagnosis, likely due to the inherent complexity of this task. To address this, we introduce Delta-KNN, a novel demonstration selection strategy that enhances ICL performance. Our method leverages a delta score to assess the relative gains of each training example, coupled with a KNN-based retriever that dynamically selects optimal "representatives" for a given input. Experiments on two AD detection datasets across three open-source LLMs demonstrate that Delta-KNN consistently outperforms existing ICL baselines. Notably, when using the Llama-3.1 model, our approach achieves new state-of-the-art results, surpassing even supervised classifiers. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:14:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Li",
"Chuyuan",
""
],
[
"Li",
"Raymond",
""
],
[
"Field",
"Thalia S.",
""
],
[
"Carenini",
"Giuseppe",
""
]
] |
2506.03477 | Masao Obata | Masao Obata, Takao Kotani and Tatsuki Oda | Efficient implementation of the quasiparticle self-consistent $GW$ method on GPU | null | null | null | null | physics.comp-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | We have developed a multi-GPU version of the quasiparticle self-consistent $GW$ (QSGW), a cutting-edge method for describing electronic excitations in a first-principles approach. While the QSGW calculation algorithm is inherently well-suited for GPU computation due to its reliance on large-scale tensor operations, achieving a maintainable and extensible implementation is not straightforward. Addressing this, we have developed a GPU version within the \texttt{ecalj} package, utilizing module-based programming style in modern Fortran. This design facilitates future development and code sustainability. Following the summary of the QSGW formalism, we present our GPU implementation approach and the results of benchmark calculations for two types of systems to demonstrate the capability of our GPU-supported QSGW calculations. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:18:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Obata",
"Masao",
""
],
[
"Kotani",
"Takao",
""
],
[
"Oda",
"Tatsuki",
""
]
] |
2506.03478 | Yuxuan Han | Yuxuan Han, Junfeng Lyu, Kuan Sheng, Minghao Que, Qixuan Zhang, Lan Xu, Feng Xu | Facial Appearance Capture at Home with Patch-Level Reflectance Prior | ACM Transactions on Graphics (Proc. of SIGGRAPH), 2025. Code: https://github.com/yxuhan/DoRA; Project Page: https://yxuhan.github.io/DoRA | null | null | null | cs.GR cs.CV | http://creativecommons.org/licenses/by/4.0/ | Existing facial appearance capture methods can reconstruct plausible facial reflectance from smartphone-recorded videos. However, the reconstruction quality is still far behind the ones based on studio recordings. This paper fills the gap by developing a novel daily-used solution with a co-located smartphone and flashlight video capture setting in a dim room. To enhance the quality, our key observation is to solve facial reflectance maps within the data distribution of studio-scanned ones. Specifically, we first learn a diffusion prior over the Light Stage scans and then steer it to produce the reflectance map that best matches the captured images. We propose to train the diffusion prior at the patch level to improve generalization ability and training stability, as current Light Stage datasets are in ultra-high resolution but limited in data size. Tailored to this prior, we propose a patch-level posterior sampling technique to sample seamless full-resolution reflectance maps from this patch-level diffusion model. Experiments demonstrate our method closes the quality gap between low-cost and studio recordings by a large margin, opening the door for everyday users to clone themselves to the digital world. Our code will be released at https://github.com/yxuhan/DoRA. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:21:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Han",
"Yuxuan",
""
],
[
"Lyu",
"Junfeng",
""
],
[
"Sheng",
"Kuan",
""
],
[
"Que",
"Minghao",
""
],
[
"Zhang",
"Qixuan",
""
],
[
"Xu",
"Lan",
""
],
[
"Xu",
"Feng",
""
]
] |
2506.03479 | Ethan Cohen | Ethan Cohen | A Real K3 Automorphism with Most of Its Entropy in the Real Part | Accompanying code in https://github.com/ethanhcoo/K3entropy | null | null | null | math.DS math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article describes an example of a real projective K3 surface admitting a real automorphism $f$ satisfying $h_{top}(f, X(\mathbb{C})) < 2 h_{top}(f, X(\mathbb{R}))$. The example presented is a $(2,2,2)$-surface in $\mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$ given by the vanishing set of $(1 + x^2)(1 + y^2)(1 + z^2) + 10xyz - 2$, first considered by McMullen. Along the way, we develop an ad hoc shadowing lemma for $C^2$ (real) surface diffeomorphisms, and apply it to estimate the location of a periodic point in $X(\mathbb{R})$. This result uses the GNU MPFR arbitrary precision arithmetic library in C and the Flipper computer program. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:23:19 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cohen",
"Ethan",
""
]
] |
2506.03480 | Seyed Amin Seyed Fakhari | Takayuki Hibi, Seyed Amin Seyed Fakhari | Bounded powers of edge ideals: The strong exchange property | null | null | null | null | math.AC math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $S=K[x_1, \ldots,x_n]$ denote the polynomial ring in $n$ variables over a field $K$ and $I \subset S$ a monomial ideal. Given a vector $\mathfrak{c}\in\mathbb{Z}_{>0}^n$, the ideal $I_{\mathfrak{c}}$ is the ideal generated by those monomials belonging to $I$ whose exponent vectors are componentwise bounded above by $\mathfrak{c}$. Let $\delta_{\mathfrak{c}}(I)$ be the largest integer $q$ for which $(I^q)_{\mathfrak{c}}\neq 0$. Let $I(G) \subset S$ denote the edge ideal of a finite graph $G$ on the vertex set $V(G) = \{x_1, \ldots, x_s\}$. In our previous work, it is shown that $(I(G)^{\delta_{\mathfrak{c}}(I)})_{\mathfrak{c}}$ is a polymatroidal ideal. Let $\mathcal{W}(\mathfrak{c},G)$ denote the minimal system of monomial generators of $(I(G)^{\delta_{\mathfrak{c}}(I)})_{\mathfrak{c}}$. It follows that $\mathcal{W}(\mathfrak{c},G)$ satisfies the symmetric exchange property. In the present paper, the question when $\mathcal{W}(\mathfrak{c},G)$ enjoys the strong exchange property, or equivalently, when $\mathcal{W}(\mathfrak{c},G)$ is of Veronese type is studied. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:35:11 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hibi",
"Takayuki",
""
],
[
"Fakhari",
"Seyed Amin Seyed",
""
]
] |
2506.03481 | Hongsong Wang | Hongsong Wang, Xiaoyan Ma, Jidong Kuang, Jie Gui | Heterogeneous Skeleton-Based Action Representation Learning | To appear in CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Skeleton-based human action recognition has received widespread attention in recent years due to its diverse range of application scenarios. Due to the different sources of human skeletons, skeleton data naturally exhibit heterogeneity. The previous works, however, overlook the heterogeneity of human skeletons and solely construct models tailored for homogeneous skeletons. This work addresses the challenge of heterogeneous skeleton-based action representation learning, specifically focusing on processing skeleton data that varies in joint dimensions and topological structures. The proposed framework comprises two primary components: heterogeneous skeleton processing and unified representation learning. The former first converts two-dimensional skeleton data into three-dimensional skeleton via an auxiliary network, and then constructs a prompted unified skeleton using skeleton-specific prompts. We also design an additional modality named semantic motion encoding to harness the semantic information within skeletons. The latter module learns a unified action representation using a shared backbone network that processes different heterogeneous skeletons. Extensive experiments on the NTU-60, NTU-120, and PKU-MMD II datasets demonstrate the effectiveness of our method in various tasks of action understanding. Our approach can be applied to action recognition in robots with different humanoid structures. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:39:13 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Hongsong",
""
],
[
"Ma",
"Xiaoyan",
""
],
[
"Kuang",
"Jidong",
""
],
[
"Gui",
"Jie",
""
]
] |
2506.03482 | Asaf Farhi Dr. | Asaf Farhi | Atomic and Molecular Waveform Processing with Attosecond Resolution | null | null | null | null | physics.optics physics.atom-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Advancing the temporal resolution in computations, signal generation and modulation, and measurements is of paramount importance for pushing the boundaries of science and technology. Optical resonators have recently demonstrated the ability to perform computational operations at frequencies beyond the gigahertz range, surpassing the speed of conventional electronic devices. However, increasing the resonator length extends the operation time but decreases the temporal resolution, with current state-of-the-art systems achieving only picosecond resolution. Here we show that atoms and molecules belong to the class of widely-used passive resonators that operate without gain, such as subwavelength particles, electric circuits, and slabs, but with long operation times and, importantly, attosecond resolution. Our analysis reveals that when resonantly exciting atoms and molecules, the resulting scattered field is the integral of the incoming field envelope, with improvement factors in temporal resolution of a million and trillion compared with optical resonators and electronic devices, respectively. We demonstrate our results theoretically for atoms and compare it with the standard slab resonator. Remarkably, our approach applies to all transition types including electronic, vibrational, rotational, and spin, with the same temporal resolution preserved across all frequencies. Our research paves the way for a new generation of devices operating on attosecond timescales and opens new avenues in fields such as computation, ultrafast phenomena, high-rate data transmission, encryption, and quantum technology. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:40:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Farhi",
"Asaf",
""
]
] |
2506.03483 | Jun Rao | Jun Rao, Zepeng Lin, Xuebo Liu, Xiaopeng Ke, Lian Lian, Dong Jin, Shengjun Cheng, Jun Yu, Min Zhang | APT: Improving Specialist LLM Performance with Weakness Case Acquisition and Iterative Preference Training | ACL2025 Findings | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) often require domain-specific fine-tuning to address targeted tasks, which risks degrading their general capabilities. Maintaining a balance between domain-specific enhancements and general model utility is a key challenge. This paper proposes a novel approach named APT (Weakness Case Acquisition and Iterative Preference Training) to enhance domain-specific performance with self-generated dis-preferred weakness data (bad cases and similar cases). APT uniquely focuses on training the model using only those samples where errors occur, alongside a small, similar set of samples retrieved for this purpose. This targeted training minimizes interference with the model's existing knowledge base, effectively retaining generic capabilities. Experimental results on the LLama-2 and Mistral-V0.3 models across various benchmarks demonstrate that APT ensures no reduction in generic capacity and achieves superior performance on downstream tasks compared to various existing methods. This validates our method as an effective strategy for enhancing domain-specific capabilities without sacrificing the model's broader applicability. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:46:38 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Rao",
"Jun",
""
],
[
"Lin",
"Zepeng",
""
],
[
"Liu",
"Xuebo",
""
],
[
"Ke",
"Xiaopeng",
""
],
[
"Lian",
"Lian",
""
],
[
"Jin",
"Dong",
""
],
[
"Cheng",
"Shengjun",
""
],
[
"Yu",
"Jun",
""
],
[
"Zhang",
"Min",
""
]
] |
2506.03484 | Melkamu Mersha | Melkamu Abay Mersha, Mesay Gemeda Yigezu, Atnafu Lambebo Tonja, Hassan Shakil, Samer Iskander, Olga Kolesnikova, Jugal Kalita | Explainable AI: XAI-Guided Context-Aware Data Augmentation | null | null | 10.1016/j.eswa.2025.128364 | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Explainable AI (XAI) has emerged as a powerful tool for improving the performance of AI models, going beyond providing model transparency and interpretability. The scarcity of labeled data remains a fundamental challenge in developing robust and generalizable AI models, particularly for low-resource languages. Conventional data augmentation techniques introduce noise, cause semantic drift, disrupt contextual coherence, lack control, and lead to overfitting. To address these challenges, we propose XAI-Guided Context-Aware Data Augmentation. This novel framework leverages XAI techniques to modify less critical features while selectively preserving most task-relevant features. Our approach integrates an iterative feedback loop, which refines augmented data over multiple augmentation cycles based on explainability-driven insights and the model performance gain. Our experimental results demonstrate that XAI-SR-BT and XAI-PR-BT improve the accuracy of models on hate speech and sentiment analysis tasks by 6.6% and 8.1%, respectively, compared to the baseline, using the Amharic dataset with the XLM-R model. XAI-SR-BT and XAI-PR-BT outperform existing augmentation techniques by 4.8% and 5%, respectively, on the same dataset and model. Overall, XAI-SR-BT and XAI-PR-BT consistently outperform both baseline and conventional augmentation techniques across all tasks and models. This study provides a more controlled, interpretable, and context-aware solution to data augmentation, addressing critical limitations of existing augmentation techniques and offering a new paradigm shift for leveraging XAI techniques to enhance AI model training. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:47:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mersha",
"Melkamu Abay",
""
],
[
"Yigezu",
"Mesay Gemeda",
""
],
[
"Tonja",
"Atnafu Lambebo",
""
],
[
"Shakil",
"Hassan",
""
],
[
"Iskander",
"Samer",
""
],
[
"Kolesnikova",
"Olga",
""
],
[
"Kalita",
"Jugal",
""
]
] |
2506.03485 | Asaf Farhi Dr. | Asaf Farhi, Dror Hershkovitz, and Haim Suchowski | Time-Domain Excitation of Complex Resonances | null | null | null | null | physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Passive resonators-systems that exhibit loss but no gain-are foundational elements across nearly every domain of physics and many types of of systems such as subwavelength particles, dielectric slabs, electric circuits, biological structures, and droplets. While their spectral properties are well characterized, their time-domain behavior under complex-frequency excitation remains largely unexplored, particularly near exceptional points where resonant modes coalesce. Here, we derive a closed-form time-domain response for passive resonators driven at complex resonance frequencies, uncovering a regime of real-frequency-like evolution at $t\ll1/\Gamma$ where they approximately function as active resonators. This framework unifies resonator behavior across disparate systems and accurately describes phenomena involving modal degeneracies and non-Hermitian effects. We verify this universality through analytical treatment of subwavelength particles and experimental demonstrations in passive electric circuits, showing excellent agreement with theory and enhanced power delivery efficiency. These results reveal a previously hidden structure in resonator dynamics and open new directions for time-domain control across across a variety of fields, including nanophotonics and circuit engineering. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:49:03 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Farhi",
"Asaf",
""
],
[
"Hershkovitz",
"Dror",
""
],
[
"Suchowski",
"Haim",
""
]
] |
2506.03486 | Tomoki Ohsawa | Tomoki Ohsawa | Explicit Symplectic Integrators for Massive Point Vortex Dynamics in Binary Mixture of Bose--Einstein Condensates | 8 pages, 5 figures | null | null | null | cond-mat.quant-gas | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct explicit integrators for massive point vortex dynamics in binary mixture of Bose--Einstein condensates proposed by Richaud et al. The integrators are symplectic and preserve the angular momentum of the system exactly. Our main focus is the small-mass regime in which the minor component of the binary mixture comprises a very small fraction of the total mass. The solution behaviors in this regime change significantly depending on the initial momenta: they are highly oscillatory unless the momenta satisfy certain conditions. The standard Runge--Kutta method performs very poorly in preserving the Hamiltonian showing a significant drift in the long run, especially for highly oscillatory solutions. On the other hand, our integrators nearly preserve the Hamiltonian without drifts. We also give an estimate of the error in the Hamiltonian by finding an asymptotic expansion of the modified Hamiltonian for our 2nd-order integrator. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 01:57:29 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ohsawa",
"Tomoki",
""
]
] |
2506.03487 | Xianming Li | Xianming Li, Aamir Shakir, Rui Huang, Julius Lipp, Jing Li | ProRank: Prompt Warmup via Reinforcement Learning for Small Language Models Reranking | null | null | null | null | cs.IR cs.CL | http://creativecommons.org/licenses/by/4.0/ | Reranking is fundamental to information retrieval and retrieval-augmented generation, with recent Large Language Models (LLMs) significantly advancing reranking quality. While recent advances with LLMs have significantly improved document reranking quality, current approaches primarily rely on large-scale LLMs (>7B parameters) through zero-shot prompting, presenting high computational costs. Small Language Models (SLMs) offer a promising alternative because of their efficiency, but our preliminary quantitative analysis reveals they struggle with understanding task prompts without fine-tuning. This limits their effectiveness for document reranking tasks. To address this issue, we introduce a novel two-stage training approach, ProRank, for SLM-based document reranking. First, we propose a prompt warmup stage using reinforcement learning GRPO to steer SLMs to understand task prompts and generate more accurate coarse-grained binary relevance scores for document reranking. Then, we continuously fine-tune the SLMs with a fine-grained score learning stage without introducing additional layers to further improve the reranking quality. Comprehensive experimental results demonstrate that the proposed ProRank consistently outperforms both the most advanced open-source and proprietary reranking models. Notably, our lightweight ProRank-0.5B model even surpasses the powerful 32B LLM reranking model on the BEIR benchmark, establishing that properly trained SLMs can achieve superior document reranking performance while maintaining computational efficiency. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:00:44 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Li",
"Xianming",
""
],
[
"Shakir",
"Aamir",
""
],
[
"Huang",
"Rui",
""
],
[
"Lipp",
"Julius",
""
],
[
"Li",
"Jing",
""
]
] |
2506.03488 | Leiming Chen | Leiming Chen, Chiu Fan Lee, John Toner | Only the Ambidextrous Can Flock: Two-dimensional Chiral Malthusian Flocks, Time crystals, and the KPZ Equation | 6 pages, 2 figures | null | null | null | cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | We study two-dimensional chiral dry Malthusian flocks; that is, chiral polar-ordered active matter with neither number nor momentum conservation. In the absence of fluctuations, these form a "time crystal", in which the velocity rotates uniformly at a fixed frequency. Fluctuations are described by the (2+1)-Kardar-Parisi-Zhang (KPZ) equation, which implies short-ranged orientational order. For weak chirality, the system is in the linear regime of the KPZ equation for a wide range of length scales, over which it exhibits quasi-long-ranged orientational order. Our predictions for velocity and density correlations are testable in both simulations and experiments. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:02:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Leiming",
""
],
[
"Lee",
"Chiu Fan",
""
],
[
"Toner",
"John",
""
]
] |
2506.03489 | Mingxu Tao | Mingxu Tao, Jie Hu, Mingchuan Yang, Yunhuai Liu, Dongyan Zhao, Yansong Feng | EpiCoDe: Boosting Model Performance Beyond Training with Extrapolation and Contrastive Decoding | ACL 2025 Findings | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The remarkable performance of Large language models (LLMs) relies heavily on the availability of abundant high-quality training data. However, the high cost of acquiring annotated data often prevents models from obtaining capabilities to tackle downstream tasks. In this paper, we introduce a novel method, EpiCoDe that boosts model performance in data-scarcity scenarios without extra training. We first employ model extrapolation to enhance a finetuned model with its inferior version, and then adopt contrastive decoding to further reduce predicted errors, by comparing the logit scores given by the extrapolated and the vanilla finetuned model. Experiments across three tasks over four different LLMs show that EpiCoDe consistently outperforms existing methods with significant and robust improvement. We also propose a new theoretical framework to reveal the mechanism behind contrastive decoding in data-scarcity scenarios, which further helps us better understand the effectiveness of EpiCoDe. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:11:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Tao",
"Mingxu",
""
],
[
"Hu",
"Jie",
""
],
[
"Yang",
"Mingchuan",
""
],
[
"Liu",
"Yunhuai",
""
],
[
"Zhao",
"Dongyan",
""
],
[
"Feng",
"Yansong",
""
]
] |
2506.03491 | Fabio Sanchez PhD | Brenda Murillo, Fabio Sanchez | Modeling Bulimia Nervosa in the Digital Age: The Role of Social Media | 6 pages, 1 figure | null | null | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | Globalization has fundamentally reshaped societal dynamics, influencing how individuals interact and perceive themselves and others. One significant consequence is the evolving landscape of eating disorders such as bulimia nervosa (BN), which are increasingly driven not just by internal psychological factors but by broader sociocultural and digital contexts. While mathematical modeling has provided valuable insights, traditional frameworks often fall short in capturing the nuanced roles of social contagion, digital media, and adaptive behavior. This review synthesizes two decades of quantitative modeling efforts, including compartmental, stochastic, and delay-based approaches. We spotlight foundational work that conceptualizes BN as a socially transmissible condition and identify critical gaps, especially regarding the intensifying impact of social media. Drawing on behavioral epidemiology and the adaptive behavior framework by Fenichel et al., we advocate for a new generation of models that incorporate feedback mechanisms, content-driven influence functions, and dynamic network effects. This work outlines a roadmap for developing more realistic, data-informed models that can guide effective public health interventions in the digital era. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:17:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Murillo",
"Brenda",
""
],
[
"Sanchez",
"Fabio",
""
]
] |
2506.03492 | Benjamin Rin | Dani\"el Andel, Benjamin Rin | Hive is PSPACE-Hard | 29 pages, 23 figures | null | null | null | cs.CC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Hive is an abstract strategy game played on a table with hexagonal pieces. First published in 2001, it was and continues to be highly popular among both casual and competitive players. In this paper, we show that for a suitably generalized version of the game, the computational problem of determining whether a given player in an arbitrary position has a winning strategy is PSPACE-hard. We do this by reduction from a variant of Generalized Geography we call Formula Game Geography. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:18:31 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Andel",
"Daniël",
""
],
[
"Rin",
"Benjamin",
""
]
] |
2506.03493 | Shiva Moshtagh | Shiva Moshtagh, Behrouz Azimian, Mohammad Golgol, Anamitra Pal | Topology-Aware Graph Neural Network-based State Estimation for PMU-Unobservable Power Systems | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by-sa/4.0/ | Traditional optimization-based techniques for time-synchronized state estimation (SE) often suffer from high online computational burden, limited phasor measurement unit (PMU) coverage, and presence of non-Gaussian measurement noise. Although conventional learning-based models have been developed to overcome these challenges, they are negatively impacted by topology changes and real-time data loss. This paper proposes a novel deep geometric learning approach based on graph neural networks (GNNs) to estimate the states of PMU-unobservable power systems. The proposed approach combines graph convolution and multi-head graph attention layers inside a customized end-to-end learning framework to handle topology changes and real-time data loss. An upper bound on SE error as a function of topology change is also derived. Experimental results for different test systems demonstrate superiority of the proposed customized GNN-SE (CGNN-SE) over traditional optimization-based techniques as well as conventional learning-based models in presence of topology changes, PMU failures, bad data, non-Gaussian measurement noise, and large system implementation. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:19:06 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Moshtagh",
"Shiva",
""
],
[
"Azimian",
"Behrouz",
""
],
[
"Golgol",
"Mohammad",
""
],
[
"Pal",
"Anamitra",
""
]
] |
2506.03494 | Jos\'e Laranjeira | Jose A. S. Laranjeira, K. A. L. Lima, Nicolas F. Martins, Luis A. Cabral, L.A. Ribeiro Junior, and Julio R. Sambrano | Propylenidene: A New Carbon Two-dimensional Material Featuring Tilted Dirac Cones | null | null | null | null | cond-mat.mtrl-sci cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Two-dimensional (2D) carbon allotropes have drawn significant interest owing to their impressive physical and chemical characteristics. Following graphene's isolation, a wide range of 2D carbon materials has been suggested, each with distinct electronic, mechanical, and optical traits. Rational design and synthesis of new 2D carbon structures hinge on experimentally reported precursors. Here, we present a 2D carbon allotrope, propylenidene (PPD), originating from the highly strained bicyclopropylidene precursor. PPD forms a rectangular lattice with 3, 8, and 10-membered carbon rings. Density functional theory (DFT) simulations investigate its structural, electronic, mechanical, and optical properties. Our study shows PPD is semi-metallic, featuring three tilted Dirac cones at the Fermi level. PPD exhibits absorption in the infrared and visible range, showing directional dependence in its response. Mechanically, PPD exhibits marked anisotropy; Young's modulus ($Y$) of 219.71 N/m in one direction and 106.16 N/m in the opposite, with an anisotropy ratio of 2.07. The shear modulus ($G$) ranges from 65.38 N/m to 32.39 N/m, yielding an anisotropy ratio of 2.02, reflecting strong directional dependence. These findings underscore the potential of this novel monolayer in applications such as energy storage, gas sensing, and optoelectronics. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:22:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Laranjeira",
"Jose A. S.",
""
],
[
"Lima",
"K. A. L.",
""
],
[
"Martins",
"Nicolas F.",
""
],
[
"Cabral",
"Luis A.",
""
],
[
"Junior",
"L. A. Ribeiro",
""
],
[
"Sambrano",
"Julio R.",
""
]
] |
2506.03495 | Jia-Hui Bi | Jia-Hui Bi, Shaoshi Yang, Sheng Chen, Ping Zhang | High-Speed Ultra-Energy-Efficient Memristor-Based Massive MIMO SIC Detector Circuit with Hybrid Analog-Digital Computing Architecture | 6 pages, 7 figures, 2 tables, to be published in IEEE Transactions on Vehicular Technology | null | 10.1109/TVT.2025.3544093 | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The emerging memristor crossbar array based computing circuits exhibit computing speeds and energy efficiency far surpassing those of traditional digital processors. This type of circuits can complete high-dimensional matrix operations in an extremely short time through analog computing, making it naturally applicable to linear detection and maximum likelihood detection in massive multiple-input multiple-output (MIMO) systems. However, the challenge of employing memristor crossbar arrays to efficiently implement other nonlinear detection algorithms, such as the successive interference cancellation (SIC) algorithm, remains unresolved. In this paper we propose a memristor-based circuit design for massive MIMO SIC detector. The proposed circuit comprises several judiciously designed analog matrix computing modules and hybrid analog-digital slicers, which enables the proposed circuit to perform the SIC algorithm with a hybrid analog-digital computing architecture. We show that the computing speed and the computational energy-efficiency of the proposed detector circuit are 43 times faster and 110 times higher, respectively, than those of a traditional 8-core digital signal processor (DSP), and also advantageous over the benchmark high-performance field programmable gate array (FPGA) and graphics processing unit (GPU). | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:22:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bi",
"Jia-Hui",
""
],
[
"Yang",
"Shaoshi",
""
],
[
"Chen",
"Sheng",
""
],
[
"Zhang",
"Ping",
""
]
] |
2506.03496 | Wee Peng Tay | Xingchao Jian and Martin G\"olz and Feng Ji and Wee Peng Tay and Abdelhak M. Zoubir | A Generalized Graph Signal Processing Framework for Multiple Hypothesis Testing over Networks | arXiv admin note: substantial text overlap with arXiv:2408.03142 | null | null | null | eess.SP cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | We consider the multiple hypothesis testing (MHT) problem over the joint domain formed by a graph and a measure space. On each sample point of this joint domain, we assign a hypothesis test and a corresponding $p$-value. The goal is to make decisions for all hypotheses simultaneously, using all available $p$-values. In practice, this problem resembles the detection problem over a sensor network during a period of time. To solve this problem, we extend the traditional two-groups model such that the prior probability of the null hypothesis and the alternative distribution of $p$-values can be inhomogeneous over the joint domain. We model the inhomogeneity via a generalized graph signal. This more flexible statistical model yields a more powerful detection strategy by leveraging the information from the joint domain. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:23:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jian",
"Xingchao",
""
],
[
"Gölz",
"Martin",
""
],
[
"Ji",
"Feng",
""
],
[
"Tay",
"Wee Peng",
""
],
[
"Zoubir",
"Abdelhak M.",
""
]
] |
2506.03497 | Kevin Wei | Oliver Guest, Kevin Wei | Bridging the Artificial Intelligence Governance Gap: The United States' and China's Divergent Approaches to Governing General-Purpose Artificial Intelligence | Published as a RAND commentary | Santa Monica, CA: RAND Corporation, 2024. https://www.rand.org/pubs/perspectives/PEA3703-1.html | 10.7249/PEA3703-1 | PE-A3703-1 | cs.CY | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The United States and China are among the world's top players in the development of advanced artificial intelligence (AI) systems, and both are keen to lead in global AI governance and development. A look at U.S. and Chinese policy landscapes reveals differences in how the two countries approach the governance of general-purpose artificial intelligence (GPAI) systems. Three areas of divergence are notable for policymakers: the focus of domestic AI regulation, key principles of domestic AI regulation, and approaches to implementing international AI governance. As AI development continues, global conversation around AI has warned of global safety and security challenges posed by GPAI systems. Cooperation between the United States and China might be needed to address these risks, and understanding the implications of these differences might help address the broader challenges for international cooperation between the United States and China on AI safety and security. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:24:27 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Guest",
"Oliver",
""
],
[
"Wei",
"Kevin",
""
]
] |
2506.03498 | Melike Gumus Akcaalan Mrs. | Onder Akcaalan, Melike Gumus Akcaalan | A Simple and Novel Passive Double-Sensitivity Optical Gyroscope Based on Non-Reciprocal Polarization Techniques | 9 pages, 4 figures, 2 tables | null | null | null | physics.app-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a novel interferometric fiber optic gyroscope (IFOG) architecture, the Double-Sensitive Non-Reciprocal Polarization Phase Shifter IFOG (DS-NRPPS-IFOG), which introduces for the first time a fully passive phase biasing scheme capable of simultaneous operation at two quadrature points $\pi/2$ and $3\pi/2$. Building upon prior passive biasing techniques, this design uses a Non-Reciprocal Polarization-Dependent Phase Shifter (NRPPS) combined with a double-pass sensing coil arrangement to achieve both passive $\pi/2$ phase modulation and enhanced measurement sensitivity. The system utilizes polarization manipulation and a quarter-wave retarder to create a double-sensitive response while eliminating the need for active modulators. Simulation results demonstrate significant performance improvements, with Angular Random Walk (ARW) values up to 50x lower than those of conventional DS-IFOG systems, depending on fiber length. Moreover, the architecture enables continuous rotation measurements and offers spontaneous noise suppression by leveraging dual quadrature detection. These findings mark a major advancement toward low-power, highly stable, and compact passive optical gyroscopes for precision navigation applications. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:24:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Akcaalan",
"Onder",
""
],
[
"Akcaalan",
"Melike Gumus",
""
]
] |
2506.03499 | Xin Tang | Xin Tang, Ruilin Li, Shuaiqi Li, Dingke Zhang | High-pressure Induced Phase Transition and Laser Characterization Response of MAPbBr$_3$ Thin Films | null | null | null | null | cond-mat.mtrl-sci physics.app-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The high-pressure behavior of 3D metal halide chalcogenides (MHPs) has been widely studied. In the field of high-pressure technology, the studies on 3D MHPs have focused on the structural and optical properties, where the optical properties are mainly investigated on the photoluminescence behavior, while the laser properties of the materials have not been studied yet. In this paper, MAPbBr$_3$-MAAc films with ionic liquid methylammonium acetate (MAAc) as solvent and conventional MAPbBr$_3$-DMF:DMSO films with N,N-dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) as solvents were prepared using solvent engineering method. In-situ pressurization testing of both materials using a small-cavity hydrostatic high-pressure device (DAC) was used to investigate the high-pressure optical behavior of the MAPbBr$_3$ films, especially the amplified spontaneous emission (ASE) properties, which, combined with high-pressure in-situ Raman, revealed that the changes in the optical properties of the films under pressure are due to the changes in the crystal structure of the materials. This paper also emphasizes that the optical properties and phase structure stability of MAPbBr$_3$-MAAc films are superior to those of MAPbBr$_3$-DMF:DMSO films under high pressure. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:26:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Tang",
"Xin",
""
],
[
"Li",
"Ruilin",
""
],
[
"Li",
"Shuaiqi",
""
],
[
"Zhang",
"Dingke",
""
]
] |
2506.03500 | Yoshitsugu Matsubara | Yoshitsugu Matsubara | Fluctuations in email size modeled using a gamma-like distribution | null | null | null | null | physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A previously established frequency distribution model, which integrates a lognormal distribution with a logarithmic equation, effectively characterizes fluctuations in email size during sending requests. In addition, an email size generation model has been developed based on this log-normal-like framework. While the fitting of these models has been deemed satisfactory, they can be further enhanced in the range of small email sizes. This study advances these models by incorporating a gamma distribution alongside a logarithmic equation. The resulting gamma-like model demonstrates a significantly improved fit compared with the log-normal-like model. These results contribute to the knowledge on the statistical properties of sending mail. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:29:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Matsubara",
"Yoshitsugu",
""
]
] |
2506.03501 | Yuchen Guo | Yuchen Guo, Zhicheng Dou, Huy H. Nguyen, Ching-Chun Chang, Saku Sugawara, Isao Echizen | Measuring Human Involvement in AI-Generated Text: A Case Study on Academic Writing | IJCNN2025 accepted | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Content creation has dramatically progressed with the rapid advancement of large language models like ChatGPT and Claude. While this progress has greatly enhanced various aspects of life and work, it has also negatively affected certain areas of society. A recent survey revealed that nearly 30% of college students use generative AI to help write academic papers and reports. Most countermeasures treat the detection of AI-generated text as a binary classification task and thus lack robustness. This approach overlooks human involvement in the generation of content even though human-machine collaboration is becoming mainstream. Besides generating entire texts, people may use machines to complete or revise texts. Such human involvement varies case by case, which makes binary classification a less than satisfactory approach. We refer to this situation as participation detection obfuscation. We propose using BERTScore as a metric to measure human involvement in the generation process and a multi-task RoBERTa-based regressor trained on a token classification task to address this problem. To evaluate the effectiveness of this approach, we simulated academic-based scenarios and created a continuous dataset reflecting various levels of human involvement. All of the existing detectors we examined failed to detect the level of human involvement on this dataset. Our method, however, succeeded (F1 score of 0.9423 and a regressor mean squared error of 0.004). Moreover, it demonstrated some generalizability across generative models. Our code is available at https://github.com/gyc-nii/CAS-CS-and-dual-head-detector | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:31:36 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Guo",
"Yuchen",
""
],
[
"Dou",
"Zhicheng",
""
],
[
"Nguyen",
"Huy H.",
""
],
[
"Chang",
"Ching-Chun",
""
],
[
"Sugawara",
"Saku",
""
],
[
"Echizen",
"Isao",
""
]
] |
2506.03502 | Yuxuan Chen | Yuxuan Chen, Haipeng Xie | CHIME: Conditional Hallucination and Integrated Multi-scale Enhancement for Time Series Diffusion Model | null | null | null | null | cs.CV cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The denoising diffusion probabilistic model has become a mainstream generative model, achieving significant success in various computer vision tasks. Recently, there has been initial exploration of applying diffusion models to time series tasks. However, existing studies still face challenges in multi-scale feature alignment and generative capabilities across different entities and long-time scales. In this paper, we propose CHIME, a conditional hallucination and integrated multi-scale enhancement framework for time series diffusion models. By employing multi-scale decomposition and adaptive integration, CHIME captures the decomposed features of time series, achieving in-domain distribution alignment between generated and original samples. In addition, we introduce a feature hallucination module in the conditional denoising process, enabling the transfer of temporal features through the training of category-independent transformation layers. Experimental results on publicly available real-world datasets demonstrate that CHIME achieves state-of-the-art performance and exhibits excellent generative generalization capabilities in few-shot scenarios. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:34:09 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Yuxuan",
""
],
[
"Xie",
"Haipeng",
""
]
] |
2506.03503 | Shan Shan | Shan Shan | Computational Architects of Society: Quantum Machine Learning for Social Rule Genesis | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | The quantification of social science remains a longstanding challenge, largely due to the philosophical nature of its foundational theories. Although quantum computing has advanced rapidly in recent years, its relevance to social theory remains underexplored. Most existing research focuses on micro-cognitive models or philosophical analogies, leaving a gap in system-level applications of quantum principles to the analysis of social systems. This study addresses that gap by proposing a theoretical and computational framework that combines quantum mechanics with Generative AI to simulate the emergence and evolution of social norms. Drawing on core quantum concepts--such as superposition, entanglement, and probabilistic measurement--this research models society as a dynamic, uncertain system and sets up five ideal-type experiments. These scenarios are simulated using 25 generative agents, each assigned evolving roles as compliers, resistors, or enforcers. Within a simulated environment monitored by a central observer (the Watcher), agents interact, respond to surveillance, and adapt to periodic normative disruptions. These interactions allow the system to self-organize under external stress and reveal emergent patterns. Key findings show that quantum principles, when integrated with generative AI, enable the modeling of uncertainty, emergence, and interdependence in complex social systems. Simulations reveal patterns including convergence toward normative order, the spread of resistance, and the spontaneous emergence of new equilibria in social rules. In conclusion, this study introduces a novel computational lens that lays the groundwork for a quantum-informed social theory. It offers interdisciplinary insights into how society can be understood not just as a structure to observe but as a dynamic system to simulate and redesign through quantum technologies. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:40:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shan",
"Shan",
""
]
] |
2506.03504 | Zhuo Zhang | Zhuo Zhuo and Xiangyu Zhang | Beyond C/C++: Probabilistic and LLM Methods for Next-Generation Software Reverse Engineering | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | This proposal discusses the growing challenges in reverse engineering modern software binaries, particularly those compiled from newer system programming languages such as Rust, Go, and Mojo. Traditional reverse engineering techniques, developed with a focus on C and C++, fall short when applied to these newer languages due to their reliance on outdated heuristics and failure to fully utilize the rich semantic information embedded in binary programs. These challenges are exacerbated by the limitations of current data-driven methods, which are susceptible to generating inaccurate results, commonly referred to as hallucinations. To overcome these limitations, we propose a novel approach that integrates probabilistic binary analysis with fine-tuned large language models (LLMs). Our method systematically models the uncertainties inherent in reverse engineering, enabling more accurate reasoning about incomplete or ambiguous information. By incorporating LLMs, we extend the analysis beyond traditional heuristics, allowing for more creative and context-aware inferences, particularly for binaries from diverse programming languages. This hybrid approach not only enhances the robustness and accuracy of reverse engineering efforts but also offers a scalable solution adaptable to the rapidly evolving landscape of software development. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:45:27 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zhuo",
"Zhuo",
""
],
[
"Zhang",
"Xiangyu",
""
]
] |
2506.03505 | Gao Xianlong | Rajamanickam Ravisankar, Sanu Kumar Gangwar, Henrique Fabrelli, Yongping Zhang, Paulsamy Muruganandam, Pankaj Kumar Mishra, Emmanuel Kengne, Gao Xianlong, and Boris A. Malomed | Stable supersolids and boselets in spin-orbit-coupled Bose-Einstein condensates with three-body interactions | 13 pages, 5 figures, to be published in Physical Review Research | null | null | null | cond-mat.quant-gas nlin.PS quant-ph | http://creativecommons.org/licenses/by/4.0/ | We explore the stability of supersolid striped waves, plane-wave boselets, and other extended states in one-dimensional spin-orbit-coupled Bose-Einstein condensates with repulsive three-body interactions (R3BIs), modeled by quintic terms in the framework of the corresponding Gross-Pitaevskii equations. In the absence of R3BIs, the extended states are susceptible to the modulational instability (MI) induced by the cubic attractive nonlinearity. Using the linearized Bogoliubov-de-Gennes equations, we identify multiple new types of MI, including baseband, passband, mixedband, and zero-wavenumber-gain ones, which give rise to deterministic rogue waves and complex nonlinear wave patterns. Our analysis reveals that R3BIs eliminate baseband and zero-wavenumber-gain MIs, forming, instead, phonon modes that enable stable boselets. Additionally, mixedband and passband MIs are suppressed, which results in a lattice-like phonon-roton mode that supports a stable supersolid phase. These stable supersolids can be realized using currently available ultracold experimental setup. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:47:46 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ravisankar",
"Rajamanickam",
""
],
[
"Gangwar",
"Sanu Kumar",
""
],
[
"Fabrelli",
"Henrique",
""
],
[
"Zhang",
"Yongping",
""
],
[
"Muruganandam",
"Paulsamy",
""
],
[
"Mishra",
"Pankaj Kumar",
""
],
[
"Kengne",
"Emmanuel",
""
],
[
"Xianlong",
"Gao",
""
],
[
"Malomed",
"Boris A.",
""
]
] |
2506.03506 | Che-Yu Chen | Ismael Ayuso, Mariam Bouhmadi-L\'opez, Che-Yu Chen, Xiao Yan Chew, Konstantinos Dialektopoulos, Yen Chin Ong | Insights in $f(Q)$ cosmology: the relevance of the connection | 21 pages, 6 figures | null | null | RIKEN-iTHEMS-Report-25 | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore the role of the affine connection in $f(Q)$ gravity, a modified theory where gravity is governed by non-metricity within the symmetric teleparallel framework. Although the connection is constrained to be flat and torsionless, it is not uniquely determined by the metric, allowing for multiple physically distinct formulations. We analyze three such connections compatible with a homogeneous and isotropic universe to show that they yield markedly different cosmological dynamics, even under the same functional form of $f(Q)$. Using both analytical and numerical methods, including a Born-Infeld type model of $f(Q)$, we demonstrate that specific connections can resolve cosmological singularities like the Big Bang and Big Rip, replacing them with smooth de Sitter phases. Others retain singularities but with notable modifications in their behavior. These findings highlight the physical relevance of connection choice in $f(Q)$ gravity and its potential to address fundamental cosmological questions. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:48:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ayuso",
"Ismael",
""
],
[
"Bouhmadi-López",
"Mariam",
""
],
[
"Chen",
"Che-Yu",
""
],
[
"Chew",
"Xiao Yan",
""
],
[
"Dialektopoulos",
"Konstantinos",
""
],
[
"Ong",
"Yen Chin",
""
]
] |
2506.03508 | Simin Wang | Tao Yu, Simin Wang, Shunqing Zhang, Xiaojing Chen, Zi Xu, Xin Wang, Jiandong Li, Junyu Liu, and Sihai Zhang | A Model-Data Dual-Driven Resource Allocation Scheme for IREE Oriented 6G Networks | Submitted to IEEE Internet of Things Journal | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The rapid and substantial fluctuations in wireless network capacity and traffic demand, driven by the emergence of 6G technologies, have exacerbated the issue of traffic-capacity mismatch, raising concerns about wireless network energy consumption. To address this challenge, we propose a model-data dual-driven resource allocation (MDDRA) algorithm aimed at maximizing the integrated relative energy efficiency (IREE) metric under dynamic traffic conditions. Unlike conventional model-driven or data-driven schemes, the proposed MDDRA framework employs a model-driven Lyapunov queue to accumulate long-term historical mismatch information and a data-driven Graph Radial bAsis Fourier (GRAF) network to predict the traffic variations under incomplete data, and hence eliminates the reliance on high-precision models and complete spatial-temporal traffic data. We establish the universal approximation property of the proposed GRAF network and provide convergence and complexity analysis for the MDDRA algorithm. Numerical experiments validate the performance gains achieved through the data-driven and model-driven components. By analyzing IREE and EE curves under diverse traffic conditions, we recommend that network operators shall spend more efforts to balance the traffic demand and the network capacity distribution to ensure the network performance, particularly in scenarios with large speed limits and higher driving visibility. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:50:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Yu",
"Tao",
""
],
[
"Wang",
"Simin",
""
],
[
"Zhang",
"Shunqing",
""
],
[
"Chen",
"Xiaojing",
""
],
[
"Xu",
"Zi",
""
],
[
"Wang",
"Xin",
""
],
[
"Li",
"Jiandong",
""
],
[
"Liu",
"Junyu",
""
],
[
"Zhang",
"Sihai",
""
]
] |
2506.03509 | Tian Shang | Tian Shang, Jianzhou Zhao, Keqi Xia, Lun-Hui Hu, Yang Xu, Qingfeng Zhan, Dariusz Jakub Gawryluk, and Toni Shiroka | Multiband superconductivity in the topological Kramers nodal-line semimetals | 9 pages, 8 figures | null | null | null | cond-mat.supr-con cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent band-structure calculations predict that the ruthenium-based ternary silicides are three-dimensional Kramers nodal line semimetals. Among them, NbRuSi and TaRuSi show bulk superconductivity (SC) below $T_c \sim 3$ K and 4 K, as well as spontaneous magnetic fields. The latter indicates the breaking of time-reversal symmetry and, thus, unconventional SC in both compounds. Previous temperature-dependent muon-spin spectroscopy studies failed to distinguish whether such compounds exhibit single-gap or multi-gap SC. Here, we report on systematic measurements of the field-dependent muon-spin relaxation rates in the superconducting state and on temperature-dependent electrical resistivity and specific heat under applied magnetic fields. Both the upper critical field and the field-dependent superconducting relaxation are well described by a two-band model. By combining our experimental results with numerical band-structure calculations, we provide solid evidence for multiband SC in NbRuSi and TaRuSi, and thus offer further insight into the unconventional- and topological nature of their superconductivity. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:50:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shang",
"Tian",
""
],
[
"Zhao",
"Jianzhou",
""
],
[
"Xia",
"Keqi",
""
],
[
"Hu",
"Lun-Hui",
""
],
[
"Xu",
"Yang",
""
],
[
"Zhan",
"Qingfeng",
""
],
[
"Gawryluk",
"Dariusz Jakub",
""
],
[
"Shiroka",
"Toni",
""
]
] |
2506.03510 | Seungcheol Park | Seungcheol Park, Sojin Lee, Jongjin Kim, Jinsik Lee, Hyunjik Jo, U Kang | Accurate Sublayer Pruning for Large Language Models by Exploiting Latency and Tunability Information | IJCAI 2025 Main Track | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | How can we accelerate large language models(LLMs) without sacrificing accuracy? The slow inference speed of LLMs hinders us to benefit from their remarkable performance in diverse applications. This is mainly because numerous sublayers are stacked together in LLMs. Sublayer pruning compresses and expedites LLMs via removing unnecessary sublayers. However, existing sublayer pruning algorithms are limited in accuracy since they naively select sublayers to prune, overlooking the different characteristics of each sublayer. In this paper, we propose SPRINT (Sublayer PRuning wIth LateNcy and Tunability Information), an accurate sublayer pruning method for LLMs. SPRINT accurately selects a target sublayer to prune by considering 1) the amount of latency reduction after pruning and 2) the tunability of sublayers. SPRINT iteratively prunes redundant sublayers and swiftly tunes the parameters of remaining sublayers. Experiments show that SPRINT achieves the best accuracy-speedup trade-off, exhibiting up to 23.88%p higher accuracy on zero-shot commonsense reasoning benchmarks compared to existing pruning algorithms. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:53:34 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Park",
"Seungcheol",
""
],
[
"Lee",
"Sojin",
""
],
[
"Kim",
"Jongjin",
""
],
[
"Lee",
"Jinsik",
""
],
[
"Jo",
"Hyunjik",
""
],
[
"Kang",
"U",
""
]
] |
2506.03511 | Fangyi Cao | Fangyi Cao, Bin Ren, Zihao Wang, Shiwei Fu, Youbin Mo, Xiaoyang Liu, Yuzhou Chen, Weixin Yao | POLARIS: A High-contrast Polarimetric Imaging Benchmark Dataset for Exoplanetary Disk Representation Learning | 9 pages main text with 5 figures, 9 pages appendix with 9 figures. Submitted to NeurIPS 2025 | null | null | null | astro-ph.EP astro-ph.IM cs.AI eess.IV | http://creativecommons.org/licenses/by/4.0/ | With over 1,000,000 images from more than 10,000 exposures using state-of-the-art high-contrast imagers (e.g., Gemini Planet Imager, VLT/SPHERE) in the search for exoplanets, can artificial intelligence (AI) serve as a transformative tool in imaging Earth-like exoplanets in the coming decade? In this paper, we introduce a benchmark and explore this question from a polarimetric image representation learning perspective. Despite extensive investments over the past decade, only a few new exoplanets have been directly imaged. Existing imaging approaches rely heavily on labor-intensive labeling of reference stars, which serve as background to extract circumstellar objects (disks or exoplanets) around target stars. With our POLARIS (POlarized Light dAta for total intensity Representation learning of direct Imaging of exoplanetary Systems) dataset, we classify reference star and circumstellar disk images using the full public SPHERE/IRDIS polarized-light archive since 2014, requiring less than 10 percent manual labeling. We evaluate a range of models including statistical, generative, and large vision-language models and provide baseline performance. We also propose an unsupervised generative representation learning framework that integrates these models, achieving superior performance and enhanced representational power. To our knowledge, this is the first uniformly reduced, high-quality exoplanet imaging dataset, rare in astrophysics and machine learning. By releasing this dataset and baselines, we aim to equip astrophysicists with new tools and engage data scientists in advancing direct exoplanet imaging, catalyzing major interdisciplinary breakthroughs. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:55:02 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cao",
"Fangyi",
""
],
[
"Ren",
"Bin",
""
],
[
"Wang",
"Zihao",
""
],
[
"Fu",
"Shiwei",
""
],
[
"Mo",
"Youbin",
""
],
[
"Liu",
"Xiaoyang",
""
],
[
"Chen",
"Yuzhou",
""
],
[
"Yao",
"Weixin",
""
]
] |
2506.03512 | Daikun Liu | Daikun Liu, Lei Cheng, Teng Wang, changyin Sun | EDCFlow: Exploring Temporally Dense Difference Maps for Event-based Optical Flow Estimation | 14 pages, 8 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent learning-based methods for event-based optical flow estimation utilize cost volumes for pixel matching but suffer from redundant computations and limited scalability to higher resolutions for flow refinement. In this work, we take advantage of the complementarity between temporally dense feature differences of adjacent event frames and cost volume and present a lightweight event-based optical flow network (EDCFlow) to achieve high-quality flow estimation at a higher resolution. Specifically, an attention-based multi-scale temporal feature difference layer is developed to capture diverse motion patterns at high resolution in a computation-efficient manner. An adaptive fusion of high-resolution difference motion features and low-resolution correlation motion features is performed to enhance motion representation and model generalization. Notably, EDCFlow can serve as a plug-and-play refinement module for RAFT-like event-based methods to enhance flow details. Extensive experiments demonstrate that EDCFlow achieves better performance with lower complexity compared to existing methods, offering superior generalization. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 02:55:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Liu",
"Daikun",
""
],
[
"Cheng",
"Lei",
""
],
[
"Wang",
"Teng",
""
],
[
"Sun",
"changyin",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.