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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.03720 | Michel Adam | Michel Adam, Patrice Frison (UBS, IRISA), Moncef Daoud (UBS), Sabine Letellier Zarshenas (UBS) | Design of a visual environment for programming by direct data manipulation | in French language | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The use of applications on computers, smartphones, and tablets has been considerably simplified thanks to interactive and dynamic graphical interfaces coupled with the mouse and touch screens. It is no longer necessary to be a computer specialist to use them. Paradoxically, the development of computer programs generally requires writing lines of code in a programming language whose syntax is particularly strict. This process poses many difficulties for programmers. We propose an original tool in which arbitrary programs (Turing-complete) can be developed in a completely visual manner by direct manipulation of the data, without writing a line of code. The user can thus develop an algorithm by directly visualizing the result of actions taken on the data. A method for constructing iterations is associated with the tool. It proposes to create each part, including the loop body, in a non-linear manner under visual control of the state of the data. In addition, the tool supports the production of lines of code in several languages including Python, C, Java, that correspond to the actions performed. In this article, we present the tool, the design choices, the problems to be solved, and the limits and the contributions of the direct-data-manipulation approach. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:50:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Adam",
"Michel",
"",
"UBS, IRISA"
],
[
"Frison",
"Patrice",
"",
"UBS, IRISA"
],
[
"Daoud",
"Moncef",
"",
"UBS"
],
[
"Zarshenas",
"Sabine Letellier",
"",
"UBS"
]
] |
2506.03721 | Florentin Millour | Florentin Millour (LAGRANGE), Guillaume Bourdarot (MPE), Jean-Baptiste Le Bouquin (IPAG), Anthony Berdeu (LIRA), Mathis Houll\'e (IPAG), Philippe Berio (LAGRANGE), Thibaut Paumard (LIRA), Denis Defr\`ere (KU Leuven), Paulo Garcia, Ferreol Soulez (CRAL), Sebastian Hoenig, Fatm\'e Allouche (ESO), Martin Bachbucher (MPE), Christophe Bailet (LAGRANGE), Cyrille Blanchard (LIRA), Olivier Boebion (LAGRANGE), Henri Bonnet (ESO), Amit Brara (MPE), Marcel Carbillet (LAGRANGE), Stephan Czempiel (MPE), Alain Delboulb\'e (IPAG), Roderick Dembet (LIRA), Cl\'emence Edouard (LIRA), Frank Eisenhauer (MPE), Halmut Feuchtgruber (MPE), Christoph Furchstsam (MPE), Stefan Gillessen (MPE), Armin Goldbrunner (MPE), Tiago Gomes, Carole Gouvret (LAGRANGE), Sylvain Guieu (IPAG), Michael Hartl (MPE), Johannes Hartwig (MPE), Frank Haussmann (MPE), David Huber (MPE), Ihsan Ibn Ta\"ieb (MPE), Johann Kolb (ESO), St\'ephane Lagarde (LAGRANGE), Olivier Lai (LAGRANGE), James Leftley (LAGRANGE), Dieter Lutz (MPE), Yves Magnard (IPAG), Hugo Nowacki (IPAG), Sylvain Oberti (ESO), Thomas Ott (MPE), Christian Rau (MPE), Sylvie Robbe-Dubois (LAGRANGE), Jules Scigliuto (LAGRANGE), Franz Soller (MPE), Pavel Shchekaturov (ESO), Daniel Schuppe (MPE), Eric Stadler (IPAG), Sinem Uysal (MPE), Felix Widmann (MPE), Ekherhard Wieprecht (MPE), Julien Woillez (ESO), Senol Yazici (MPE) | GRAVITY+ adaptive optics (GPAO) tests in Europe | null | Optical and Infrared Interferometry and Imaging IX, Jun 2024, Yokohama, France. pp.72 | 27E8;10.1117/12.3020269 | null | astro-ph.IM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present in this proceeding the results of the test phase of the GRAVITY+ adaptive optics. This extreme AO will enable both high-dynamic range observations of faint companions (including exoplanets) thanks to a 40x40 sub-apertures wavefront control, and sensitive observations (including AGNs) thanks to the addition of a laser guide star to each UT of the VLT. This leap forward is made thanks to a mostly automated setup of the AO, including calibration of the NCPAs, that we tested in Europe on the UT+atmosphere simulator we built in Nice. We managed to reproduce in laboratory the expected performances of all the modes of the AO, including under non-optimal atmospheric or telescope alignment conditions, giving us the green light to proceed with the Assembly, Integration and Verification phase in Paranal. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:53:30 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Millour",
"Florentin",
"",
"LAGRANGE"
],
[
"Bourdarot",
"Guillaume",
"",
"MPE"
],
[
"Bouquin",
"Jean-Baptiste Le",
"",
"IPAG"
],
[
"Berdeu",
"Anthony",
"",
"LIRA"
],
[
"Houllé",
"Mathis",
"",
"IPAG"
],
[
"Berio",
"Philippe",
"",
"LAGRANGE"
],
[
"Paumard",
"Thibaut",
"",
"LIRA"
],
[
"Defrère",
"Denis",
"",
"KU Leuven"
],
[
"Garcia",
"Paulo",
"",
"CRAL"
],
[
"Soulez",
"Ferreol",
"",
"CRAL"
],
[
"Hoenig",
"Sebastian",
"",
"ESO"
],
[
"Allouche",
"Fatmé",
"",
"ESO"
],
[
"Bachbucher",
"Martin",
"",
"MPE"
],
[
"Bailet",
"Christophe",
"",
"LAGRANGE"
],
[
"Blanchard",
"Cyrille",
"",
"LIRA"
],
[
"Boebion",
"Olivier",
"",
"LAGRANGE"
],
[
"Bonnet",
"Henri",
"",
"ESO"
],
[
"Brara",
"Amit",
"",
"MPE"
],
[
"Carbillet",
"Marcel",
"",
"LAGRANGE"
],
[
"Czempiel",
"Stephan",
"",
"MPE"
],
[
"Delboulbé",
"Alain",
"",
"IPAG"
],
[
"Dembet",
"Roderick",
"",
"LIRA"
],
[
"Edouard",
"Clémence",
"",
"LIRA"
],
[
"Eisenhauer",
"Frank",
"",
"MPE"
],
[
"Feuchtgruber",
"Halmut",
"",
"MPE"
],
[
"Furchstsam",
"Christoph",
"",
"MPE"
],
[
"Gillessen",
"Stefan",
"",
"MPE"
],
[
"Goldbrunner",
"Armin",
"",
"MPE"
],
[
"Gomes",
"Tiago",
"",
"LAGRANGE"
],
[
"Gouvret",
"Carole",
"",
"LAGRANGE"
],
[
"Guieu",
"Sylvain",
"",
"IPAG"
],
[
"Hartl",
"Michael",
"",
"MPE"
],
[
"Hartwig",
"Johannes",
"",
"MPE"
],
[
"Haussmann",
"Frank",
"",
"MPE"
],
[
"Huber",
"David",
"",
"MPE"
],
[
"Taïeb",
"Ihsan Ibn",
"",
"MPE"
],
[
"Kolb",
"Johann",
"",
"ESO"
],
[
"Lagarde",
"Stéphane",
"",
"LAGRANGE"
],
[
"Lai",
"Olivier",
"",
"LAGRANGE"
],
[
"Leftley",
"James",
"",
"LAGRANGE"
],
[
"Lutz",
"Dieter",
"",
"MPE"
],
[
"Magnard",
"Yves",
"",
"IPAG"
],
[
"Nowacki",
"Hugo",
"",
"IPAG"
],
[
"Oberti",
"Sylvain",
"",
"ESO"
],
[
"Ott",
"Thomas",
"",
"MPE"
],
[
"Rau",
"Christian",
"",
"MPE"
],
[
"Robbe-Dubois",
"Sylvie",
"",
"LAGRANGE"
],
[
"Scigliuto",
"Jules",
"",
"LAGRANGE"
],
[
"Soller",
"Franz",
"",
"MPE"
],
[
"Shchekaturov",
"Pavel",
"",
"ESO"
],
[
"Schuppe",
"Daniel",
"",
"MPE"
],
[
"Stadler",
"Eric",
"",
"IPAG"
],
[
"Uysal",
"Sinem",
"",
"MPE"
],
[
"Widmann",
"Felix",
"",
"MPE"
],
[
"Wieprecht",
"Ekherhard",
"",
"MPE"
],
[
"Woillez",
"Julien",
"",
"ESO"
],
[
"Yazici",
"Senol",
"",
"MPE"
]
] |
2506.03722 | Huiyan Li | Yinfeng Xia and Huiyan Li and Chenyang Le and Manhong Wang and Yutao Sun and Xingyang Ma and Yanmin Qian | MFLA: Monotonic Finite Look-ahead Attention for Streaming Speech Recognition | Accepted by Interspeech 2025 | null | null | null | cs.CL cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | Applying large pre-trained speech models like Whisper has shown promise in reducing training costs for various speech tasks. However, integrating these models into streaming systems remains a challenge. This paper presents a novel prefix-to-prefix training framework for streaming recognition by fine-tuning the Whisper. We introduce the Continuous Integrate-and-Fire mechanism to establish a quasi-monotonic alignment between continuous speech sequences and discrete text tokens. Additionally, we design Monotonic Finite Look-ahead Attention, allowing each token to attend to infinite left-context and finite right-context from the speech sequences. We also employ the wait-k decoding strategy to simplify the decoding process while ensuring consistency between training and testing. Our theoretical analysis and experiments demonstrate that this approach achieves a controllable trade-off between latency and quality, making it suitable for various streaming applications. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:53:40 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xia",
"Yinfeng",
""
],
[
"Li",
"Huiyan",
""
],
[
"Le",
"Chenyang",
""
],
[
"Wang",
"Manhong",
""
],
[
"Sun",
"Yutao",
""
],
[
"Ma",
"Xingyang",
""
],
[
"Qian",
"Yanmin",
""
]
] |
2506.03723 | Chaeyun Jang | Chaeyun Jang, Moonseok Choi, Yegon Kim, Hyungi Lee, Juho Lee | Verbalized Confidence Triggers Self-Verification: Emergent Behavior Without Explicit Reasoning Supervision | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Uncertainty calibration is essential for the safe deployment of large language models (LLMs), particularly when users rely on verbalized confidence estimates. While prior work has focused on classifiers or short-form generation, confidence calibration for chain-of-thought (CoT) reasoning remains largely unexplored. Surprisingly, we find that supervised fine-tuning with scalar confidence labels alone suffices to elicit self-verification behavior of language models, without any explicit reasoning supervision or reinforcement learning-based rewards. Despite being trained only to produce a verbalized confidence score without any self-verifying examples, the model learns to generate longer and self-checking responses for low-confidence queries while providing more concise answers for high-confidence ones. We further propose a simple rethinking method that boosts performance via test-time scaling based on calibrated uncertainty. Experiments on GSM8K and held-out reasoning tasks such as MATH-500 and ARC-Challenge show that our confidence-aware fine-tuning improves both calibration and accuracy, while also enhancing interpretability by aligning the model's reasoning path with its confidence. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:56:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jang",
"Chaeyun",
""
],
[
"Choi",
"Moonseok",
""
],
[
"Kim",
"Yegon",
""
],
[
"Lee",
"Hyungi",
""
],
[
"Lee",
"Juho",
""
]
] |
2506.03724 | Weixiong Mai | Ping Liang, Pei Dang, Weixiong Mai | Uncertainty principles for free metaplectic transformation and associated metaplectic operators | 42 pages | null | null | null | math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we systematically investigate the Heisenberg-Pauli-Weyl uncertainty principle for free metaplectic transformation, as well as metaplectic operators. Specifically, we obtain two different types of the uncertainty principle for free metaplectic transformations in terms of the so-called phase derivative, one of which can be generalized to the $L^p$-case with $1\le p\le 2$. The obtained results are valid not only for free metaplectic transformations but also for general metaplectic operators. In particular, we point out that our results are closely related to those given in \cite{Dias-deGosson-Prata}, and the relationship should be new and not exactly given in the existing literature. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:58:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Liang",
"Ping",
""
],
[
"Dang",
"Pei",
""
],
[
"Mai",
"Weixiong",
""
]
] |
2506.03725 | Daniil Medyakov Mr. | Daniil Medyakov, Sergey Stanko, Gleb Molodtsov, Philip Zmushko, Grigoriy Evseev, Egor Petrov, Aleksandr Beznosikov | Sign-SGD is the Golden Gate between Multi-Node to Single-Node Learning: Significant Boost via Parameter-Free Optimization | 58 pages, 5 figures, 5 tables | null | null | null | cs.LG math.OC | http://creativecommons.org/licenses/by/4.0/ | Quite recently, large language models have made a significant breakthrough across various disciplines. However, training them is an extremely resource-intensive task, even for major players with vast computing resources. One of the methods gaining popularity in light of these challenges is Sign-SGD. This method can be applied both as a memory-efficient approach in single-node training and as a gradient compression technique in the distributed learning. Nevertheless, it is impossible to automatically determine the effective stepsize from the theoretical standpoint. Indeed, it depends on the parameters of the dataset to which we do not have access in the real-world learning paradigm. To address this issue, we design several variants of single-node deterministic Sign-SGD. We extend our approaches to practical scenarios: stochastic single-node and multi-node learning, methods with incorporated momentum. We conduct extensive experiments on real machine learning problems that emphasize the practical applicability of our ideas. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:58:33 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Medyakov",
"Daniil",
""
],
[
"Stanko",
"Sergey",
""
],
[
"Molodtsov",
"Gleb",
""
],
[
"Zmushko",
"Philip",
""
],
[
"Evseev",
"Grigoriy",
""
],
[
"Petrov",
"Egor",
""
],
[
"Beznosikov",
"Aleksandr",
""
]
] |
2506.03726 | Christian Leibel | Christian Leibel and Lutz Bornmann | Introducing multiverse analysis to bibliometrics: The case of team size effects on disruptive research | 50 pages, 3 figures, 11 tables | null | null | null | cs.DL stat.AP | http://creativecommons.org/licenses/by/4.0/ | Although bibliometrics has become an essential tool in the evaluation of research performance, bibliometric analyses are sensitive to a range of methodological choices. Subtle choices in data selection, indicator construction, and modeling decisions can substantially alter results. Ensuring robustness, meaning that findings hold up under different reasonable scenarios, is therefore critical for credible research and research evaluation. To address this issue, this study introduces multiverse analysis to bibliometrics. Multiverse analysis is a statistical tool that enables analysts to transparently discuss modeling assumptions and thoroughly assess model robustness. Whereas standard robustness checks usually cover only a small subset of all plausible models, multiverse analysis includes all plausible models. We illustrate the benefits of multiverse analysis by testing the hypothesis posed by Wu et al. (2019) that small teams produce more disruptive research than large teams. While we found robust evidence of a negative effect of team size on disruption scores, the effect size is so small that its practical relevance seems questionable. Our findings underscore the importance of assessing the multiverse robustness of bibliometric results to clarify their practical implications. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 08:58:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Leibel",
"Christian",
""
],
[
"Bornmann",
"Lutz",
""
]
] |
2506.03727 | Konstantin Borovkov | Aaron Chong and Konstantin Borovkov | Large deviation probabilities for sums of censored random variables with regularly varying distribution tails | 18 pages, 2 figures | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $\xi_1, \xi_2,\ldots$ be a sequence of independent and identically distributed random variables with zero mean, finite second moment and regularly varying right distribution tail. Motivated by a stop-loss insurance model, we consider a threshold sequence $M_n\gg (n\ln n)^{1/2},$ $n\to \infty,$ and establish the asymptotics of the probabilities of the large deviations of the form $ \sum_{j=1}^n(\xi_j \wedge M_n)>x$ in the whole spectrum of $x$-values in the region $O(M_n).$ The asymptotic representations for these probabilities obey the "multiple large jumps principle" and have different forms in the vicinities of the multiples $kM_n$ of the censoring threshold values, on the one hand, and inside intervals of the form $((k-1)M_n, kM_n),$ on the other. We show that there is a "smooth transition" of these representations from one to the other when the deviation $x$ increases to a multiple of $M_n$, "crosses" it and then moves away from it. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:00:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chong",
"Aaron",
""
],
[
"Borovkov",
"Konstantin",
""
]
] |
2506.03728 | Hang Fan | Hang Fan, Mingxuan Li, Jingshi Cui, Zuhan Zhang, Wencai Run and Dunnan Liu | Spatiotemporal Prediction of Electric Vehicle Charging Load Based on Large Language Models | null | null | null | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | The rapid growth of EVs and the subsequent increase in charging demand pose significant challenges for load grid scheduling and the operation of EV charging stations. Effectively harnessing the spatiotemporal correlations among EV charging stations to improve forecasting accuracy is complex. To tackle these challenges, we propose EV-LLM for EV charging loads based on LLMs in this paper. EV-LLM integrates the strengths of Graph Convolutional Networks (GCNs) in spatiotemporal feature extraction with the generalization capabilities of fine-tuned generative LLMs. Also, EV-LLM enables effective data mining and feature extraction across multimodal and multidimensional datasets, incorporating historical charging data, weather information, and relevant textual descriptions to enhance forecasting accuracy for multiple charging stations. We validate the effectiveness of EV-LLM by using charging data from 10 stations in California, demonstrating its superiority over the other traditional deep learning methods and potential to optimize load grid scheduling and support vehicle-to-grid interactions. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:03:30 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fan",
"Hang",
""
],
[
"Li",
"Mingxuan",
""
],
[
"Cui",
"Jingshi",
""
],
[
"Zhang",
"Zuhan",
""
],
[
"Run",
"Wencai",
""
],
[
"Liu",
"Dunnan",
""
]
] |
2506.03729 | Shailendra Bhandari | Shailendra Bhandari, Pedro Lencastre, Sergiy Denysov, Yurii Bystryk, Pedro G. Lind | IntLevPy: A Python library to classify and model intermittent and L\'evy processes | 6 pages, 2 figures | null | null | null | cs.NE cs.MS | http://creativecommons.org/licenses/by/4.0/ | IntLevPy provides a comprehensive description of the IntLevPy Package, a Python library designed for simulating and analyzing intermittent and L\'evy processes. The package includes functionalities for process simulation, including full parameter estimation and fitting optimization for both families of processes, moment calculation, and classification methods. The classification methodology utilizes adjusted-$R^2$ and a noble performance measure {\Gamma}, enabling the distinction between intermittent and L\'evy processes. IntLevPy integrates iterative parameter optimization with simulation-based validation. This paper provides an in-depth user guide covering IntLevPy software architecture, installation, validation workflows, and usage examples. In this way, IntLevPy facilitates systematic exploration of these two broad classes of stochastic processes, bridging theoretical models and practical applications. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:03:58 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bhandari",
"Shailendra",
""
],
[
"Lencastre",
"Pedro",
""
],
[
"Denysov",
"Sergiy",
""
],
[
"Bystryk",
"Yurii",
""
],
[
"Lind",
"Pedro G.",
""
]
] |
2506.03730 | Yigit Sozen | Yigit Sozen, Yu Kyoung Ryu, Javier Martinez, and Andres Castellanos-Gomez | Direct laser ablation of 2D material films for fabricating multi-functional flexible and transparent devices | 31 pages, 7 figures | Advanced Functional Materials, 2507458 (2025) | 10.1002/adfm.202507458 | null | physics.app-ph cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present a scalable method for direct patterning of graphite and transition metal dichalcogenide (TMD) films on polycarbonate (PC) and other transparent substrates using fiber laser ablation. This process facilitates the fabrication of various functional devices, including strain gauges, supercapacitors, and photodetector arrays, without the need for photolithography or solvents, thereby simplifying device production and enhancing environmental sustainability. Utilizing roll-to-roll mechanical exfoliation, homogeneous nanosheet films are created and then patterned with a laser engraving system. Electrical and optical characterization confirms that the laser-processed films maintain their crystallinity, with no observable damage to the underlying substrate. We demonstrate the scalability of this approach by constructing a WSe2/graphite photodetector array on PC, which exhibits high sensitivity, low noise, and uniform photocurrent response across its active channels. As a proof-of-concept, this array is used as an image sensor to capture light patterns, showcasing its potential for flexible and semi-transparent imaging applications. These findings open up new avenues for incorporating all-van der Waals devices into wearable electronics, optoelectronics, and imaging technologies. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:05:12 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sozen",
"Yigit",
""
],
[
"Ryu",
"Yu Kyoung",
""
],
[
"Martinez",
"Javier",
""
],
[
"Castellanos-Gomez",
"Andres",
""
]
] |
2506.03731 | Zhengyang Li | Zhengyang Li | Enhancing Text Comprehension for Dyslexic Readers: A 3D Semantic Visualization Approach Using Transformer Mode | null | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dyslexic individuals often face significant challenges with traditional reading, particularly when engaging with complex texts such as mystery novels. These texts typically demand advanced narrative tracking and information integration skills, making it difficult for dyslexic readers to fully comprehend the content. However, research indicates that while dyslexic individuals may struggle with textual processing, they often possess strong spatial imagination abilities. Leveraging this strength, this study proposes an innovative approach using Transformer models to map sentences and words into three-dimensional vector representations. This process clusters semantically similar sentences and words in spatial proximity, allowing dyslexic readers to interpret the semantic structure and narrative flow of the text through spatial perception. Experimental results demonstrate that, compared to direct text reading, this three-dimensional semantic visualization method significantly enhances dyslexic readers' comprehension of complex texts. In particular, it shows marked advantages in identifying narrative relationships and character connections. This study provides a novel pathway for improving textual comprehension among dyslexic individuals | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:05:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Li",
"Zhengyang",
""
]
] |
2506.03732 | Nikolaos Mavredakis Dr | N. Mavredakis, A. Pacheco-Sanchez, R. Garcia Cortadella, Anton-Guimer\`a-Brunet, J. A. Garrido and D. Jim\'enez | Physics-Based Compact Modeling for the Drain Current Variability in Single-Layer Graphene FETs | null | IEEE Trans. Electron Devices, vol. 72, no. 6, pp. 3314-3321, Jun. 2025 | 10.1109/TED.2025.3560616 | null | cond-mat.mes-hall physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | For the growth of emerging graphene field-effect transistor (GFET) technologies, a thorough characterization of on-wafer variability is required. Here, we report for the first time a physics-based compact model, which precisely describes the drain current (ID) fluctuations of monolayer GFETs. Physical mechanisms known to generate 1/f noise in transistors, such as carrier number and Coulomb scattering mobility fluctuations, are also revealed to cause ID variance. Such effects are considered in the model by being activated locally in the channel and the integration of their contributions from source to drain results in total variance. The proposed model is experimentally validated from a statistical population of three different-sized solution-gated (SG) GFETs from strong p- to strong n-type bias conditions. A series resistance ID variance model is also derived mainly contributing at high carrier densities. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:05:47 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mavredakis",
"N.",
""
],
[
"Pacheco-Sanchez",
"A.",
""
],
[
"Cortadella",
"R. Garcia",
""
],
[
"Anton-Guimerà-Brunet",
"",
""
],
[
"Garrido",
"J. A.",
""
],
[
"Jiménez",
"D.",
""
]
] |
2506.03733 | Kyung Hoon Han | Kyung Hoon Han, Seung-Hyeok Kye | Supporting hyperplanes for Schmidt numbers and Schmidt number witnesses | 12 pages | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the compact convex set of all bi-partite states of Schmidt number less than or equal to $k$, together with that of $k$-blockpositive matrices of trace one, which play the roles of Schmidt number witnesses. In this note, we look for hyperplanes which support those convex sets and are perpendicular to a one parameter family through the maximally mixed state. We show that this is equivalent to determining the intervals for the dual objects on the one parameter family. We illustrate our results for the one parameter families including Werner states and isotropic states. Through the discussion, we give a simple decomposition of the separable Werner state into the sum of product states. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:06:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Han",
"Kyung Hoon",
""
],
[
"Kye",
"Seung-Hyeok",
""
]
] |
2506.03734 | Aaron Poole | Imtak Jeon, Hyojoong Kim, Nakwoo Kim, Aaron Poole and Augniva Ray | Supersymmetric localisation of $\mathcal{N}=(2,2)$ theories on a spindle | 47 pages + appendices | null | null | null | hep-th | http://creativecommons.org/licenses/by/4.0/ | We consider two-dimensional $\mathcal{N}=(2,2)$ supersymmteric field theories living on a weighted projective space $\mathbb{WCP}_{[n_1,n_2]}^1$, often referred to as a spindle. Starting from the spindle solution of five-dimensional minimal gauged supergravity, we construct a two-dimensional $\mathcal{N}=(2,2)$ theory on a spindle, preserving supersymmetry via the anti-twist mechanism and admitting two Killing spinors of opposing $R$-charge. We apply the technique of supersymmetric localisation to compute the exact partition function for a theory consisting of an abelian vector multiplet and a chiral multiplet, finding that the path integral localises to a real moduli space of vector multiplet fluctuations. We compute the one-loop determinants via the equivariant index, using both the method of unpaired eigenvalues and the fixed point theorem, finding agreement between the two approaches. We conclude with the explicit example of a charged chiral multiplet in the presence of a Fayet-Iliopoulos term. This work paves the way towards uncovering two-dimensional dualities, such as mirror symmetry, for field theories defined on orbifold backgrounds. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:06:33 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Jeon",
"Imtak",
""
],
[
"Kim",
"Hyojoong",
""
],
[
"Kim",
"Nakwoo",
""
],
[
"Poole",
"Aaron",
""
],
[
"Ray",
"Augniva",
""
]
] |
2506.03735 | Junling Wang | Junling Wang, Anna Rutkiewicz, April Yi Wang, Mrinmaya Sachan | Generating Pedagogically Meaningful Visuals for Math Word Problems: A New Benchmark and Analysis of Text-to-Image Models | Findings of the Association for Computational Linguistics: ACL 2025 | null | null | null | cs.CL cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visuals are valuable tools for teaching math word problems (MWPs), helping young learners interpret textual descriptions into mathematical expressions before solving them. However, creating such visuals is labor-intensive and there is a lack of automated methods to support this process. In this paper, we present Math2Visual, an automatic framework for generating pedagogically meaningful visuals from MWP text descriptions. Math2Visual leverages a pre-defined visual language and a design space grounded in interviews with math teachers, to illustrate the core mathematical relationships in MWPs. Using Math2Visual, we construct an annotated dataset of 1,903 visuals and evaluate Text-to-Image (TTI) models for their ability to generate visuals that align with our design. We further fine-tune several TTI models with our dataset, demonstrating improvements in educational visual generation. Our work establishes a new benchmark for automated generation of pedagogically meaningful visuals and offers insights into key challenges in producing multimodal educational content, such as the misrepresentation of mathematical relationships and the omission of essential visual elements. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:08:11 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Junling",
""
],
[
"Rutkiewicz",
"Anna",
""
],
[
"Wang",
"April Yi",
""
],
[
"Sachan",
"Mrinmaya",
""
]
] |
2506.03736 | Vladimir Zykunov | Vladimir Zykunov | Influence of photon inverse emission on forward-backward asymmetry in dilepton production at LHC | 10 pages, 10 figures, 1 table | null | null | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | Contribution of photon inverse emission to the dilepton production process in hadron collisions for Large Hadron Collider (LHC) experimental program is calculated in details. Numerical analysis of inverse emission effects on observable quantities (cross sections and forward-backward asymmetry) is performed in a wide kinematical region including the CMS LHC experiment in Run3/HL regime which corresponds to ultra-high energies and dilepton invariant masses. Effective technics of additive relative corrections is applied for analysis of the influence of different radiative contributions on forward-backward asymmetry. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:08:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zykunov",
"Vladimir",
""
]
] |
2506.03737 | Hao Yu | Hao Yu, Tangyu Jiang, Shuning Jia, Shannan Yan, Shunning Liu, Haolong Qian, Guanghao Li, Shuting Dong, Huaisong Zhang, Chun Yuan | ComRoPE: Scalable and Robust Rotary Position Embedding Parameterized by Trainable Commuting Angle Matrices | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | The Transformer architecture has revolutionized various regions since it was proposed, and its effectiveness largely depends on the ability to encode positional information. Traditional position encoding methods exhibit significant limitations due to lack of robustness and flexibility of position. Therefore, Rotary Positional Encoding (RoPE) was proposed to alleviate these issues, which integrates positional information by rotating the embeddings in the attention mechanism. However, RoPE requires manually defined rotation matrices with limited transformation space, constraining the model's capacity. In this work, we propose ComRoPE, which generalizes RoPE by defining it in terms of trainable commuting angle matrices. Specifically, we demonstrate that pairwise commutativity of these matrices is essential for RoPE to achieve scalability and positional robustness. We formally define the RoPE Equation, which is an essential condition that ensures consistent performance with position offsets. Based on the theoretical analysis, we present two types of trainable commuting angle matrices as sufficient solutions to the RoPE equation, which significantly improve performance, surpassing the current state-of-the-art method by 1.6% at training resolution and 2.9% at higher resolution on the ImageNet-1K dataset. Furthermore, our framework shows versatility in generalizing to existing RoPE formulations and offering new insights for future positional encoding research. To ensure reproducibility, the source code and instructions are available at https://github.com/Longin-Yu/ComRoPE | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:10:02 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Yu",
"Hao",
""
],
[
"Jiang",
"Tangyu",
""
],
[
"Jia",
"Shuning",
""
],
[
"Yan",
"Shannan",
""
],
[
"Liu",
"Shunning",
""
],
[
"Qian",
"Haolong",
""
],
[
"Li",
"Guanghao",
""
],
[
"Dong",
"Shuting",
""
],
[
"Zhang",
"Huaisong",
""
],
[
"Yuan",
"Chun",
""
]
] |
2506.03738 | Vittorio Bianco | Francesca Borrelli, Giusy Giugliano, Emilie Houliez, Jaromir Behal, Daniele Pirone, Leonilde Roselli, Angela Sardo, Valerio Zupo, Maria Costantini, Lisa Miccio, Pasquale Memmolo, Vittorio Bianco, Pietro Ferraro | 3D Holographic Flow Cytometry Measurements of Microalgae: Strategies for Angle Recovery in Complex Rotation Patterns | null | null | null | null | physics.bio-ph cs.SY eess.IV eess.SY | http://creativecommons.org/licenses/by/4.0/ | Marine ecosystems are in the spotlight, because environmental changes are threatening biodiversity and ecological functions. In this context, microalgae play key ecological roles both in planktonic and benthic ecosystems. Consequently, they are considered indispensable targets for global monitoring programs. However, due to a high spatial and temporal variability and to difficulties of species identification (still relying on microscopy observations), the assessment of roles played by these components of marine ecosystems is demanding. In addition, technologies for a 3D assessment of their complex morphology are scarcely available. Here, we present a comprehensive workflow for retrieving 3D information on microalgae with diverse geometries through holographic microscopy operating in flow-cytometry mode. Depending on the rotation patterns of samples, a tailored approach is used to retrieve their rolling angles. We demonstrate the feasibility of measuring 3D data of various microalgae, contingent to the intrinsic optical properties of cells. Specifically, we show that for quasi-transparent and low-scattering microorganisms, the retrieved angles permit to achieve quantitative 3D tomographic Refractive Index (RI) mapping, providing a full characterization of the alga in terms of its inner structure and the outer shape. Moreover, even in the most challenging scenarios, where microalgae exhibit high light absorption or strong scattering, quantitative 3D shape reconstructions of diatoms and dinoflagellates can be at least achieved. Finally, we compare our direct 3D measurements with 2D inferences of 3D properties, obtained using a commercially available microscopy system. The ability to non-invasively obtain 3D information on microalgae marks a fundamental advancement in the field, unlocking a wealth of novel biological insights for characterizing aquatic ecosystems. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:10:16 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Borrelli",
"Francesca",
""
],
[
"Giugliano",
"Giusy",
""
],
[
"Houliez",
"Emilie",
""
],
[
"Behal",
"Jaromir",
""
],
[
"Pirone",
"Daniele",
""
],
[
"Roselli",
"Leonilde",
""
],
[
"Sardo",
"Angela",
""
],
[
"Zupo",
"Valerio",
""
],
[
"Costantini",
"Maria",
""
],
[
"Miccio",
"Lisa",
""
],
[
"Memmolo",
"Pasquale",
""
],
[
"Bianco",
"Vittorio",
""
],
[
"Ferraro",
"Pietro",
""
]
] |
2506.03739 | Vincent Couderc Dr | Yago Arosa, Tigran Mansuryan, Arnaud Poisson, Wasyhun Asefa Gemechu, Katarzyna Krupa, Mario Ferraro, Fabio Mangini, Benjamin Wetzel, Stefan Wabnitz, Alessandro Tonello, Vincent Couderc | Spatio-spectral light-by-light moulding in multimode fibre | null | null | null | null | physics.optics nlin.AO | http://creativecommons.org/licenses/by/4.0/ | Controlling complex light waves to achieve desired behaviours or characteristics on demand presents a significant challenge. This task becomes even more complicated when manipulating speckled light beams owing to their inherently fuzzy intensity and phase structures. Here, we demonstrate that a weak speckled second-harmonic signal in a multimode graded-index fibre can be manipulated via its conservative interaction with a high-power co-propagating fundamental pump wave. Specifically, the spatial quality of the signal can be either enhanced or degraded by varying the pump's power or its modal power distribution. The underlying physical mechanism is the optically induced mode conversion, whose phase-matching can be controlled by the mode power distribution of the pump beam. This phenomenon enables new possibilities for manipulating complex light via material nonlinearities in multimode guiding structures. A striking example of this novel light-by-light control is the experimentally observed enhancement or partial suppression of the visible Raman Stokes cascade regulated by the second harmonic beam, while modulated by the mode power distribution of the fundamental beam. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:11:33 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Arosa",
"Yago",
""
],
[
"Mansuryan",
"Tigran",
""
],
[
"Poisson",
"Arnaud",
""
],
[
"Gemechu",
"Wasyhun Asefa",
""
],
[
"Krupa",
"Katarzyna",
""
],
[
"Ferraro",
"Mario",
""
],
[
"Mangini",
"Fabio",
""
],
[
"Wetzel",
"Benjamin",
""
],
[
"Wabnitz",
"Stefan",
""
],
[
"Tonello",
"Alessandro",
""
],
[
"Couderc",
"Vincent",
""
]
] |
2506.03740 | Jianfeng Wu | Jianfeng Wu, Nannan Xu | SAAT: Synergistic Alternating Aggregation Transformer for Image Super-Resolution | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Single image super-resolution is a well-known downstream task which aims to restore low-resolution images into high-resolution images. At present, models based on Transformers have shone brightly in the field of super-resolution due to their ability to capture long-term dependencies in information. However, current methods typically compute self-attention in nonoverlapping windows to save computational costs, and the standard self-attention computation only focuses on its results, thereby neglecting the useful information across channels and the rich spatial structural information generated in the intermediate process. Channel attention and spatial attention have, respectively, brought significant improvements to various downstream visual tasks in terms of extracting feature dependency and spatial structure relationships, but the synergistic relationship between channel and spatial attention has not been fully explored yet.To address these issues, we propose a novel model. Synergistic Alternating Aggregation Transformer (SAAT), which can better utilize the potential information of features. In SAAT, we introduce the Efficient Channel & Window Synergistic Attention Group (CWSAG) and the Spatial & Window Synergistic Attention Group (SWSAG). On the one hand, CWSAG combines efficient channel attention with shifted window attention, enhancing non-local feature fusion, and producing more visually appealing results. On the other hand, SWSAG leverages spatial attention to capture rich structured feature information, thereby enabling SAAT to more effectively extract structural features.Extensive experimental results and ablation studies demonstrate the effectiveness of SAAT in the field of super-resolution. SAAT achieves performance comparable to that of the state-of-the-art (SOTA) under the same quantity of parameters. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:12:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wu",
"Jianfeng",
""
],
[
"Xu",
"Nannan",
""
]
] |
2506.03741 | Daniel Buschek | Rifat Mehreen Amin, Oliver Hans K\"uhle, Daniel Buschek, Andreas Butz | PromptCanvas: Composable Prompting Workspaces Using Dynamic Widgets for Exploration and Iteration in Creative Writing | null | null | null | null | cs.HC cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce PromptCanvas, a concept that transforms prompting into a composable, widget-based experience on an infinite canvas. Users can generate, customize, and arrange interactive widgets representing various facets of their text, offering greater control over AI-generated content. PromptCanvas allows widget creation through system suggestions, user prompts, or manual input, providing a flexible environment tailored to individual needs. This enables deeper engagement with the creative process. In a lab study with 18 participants, PromptCanvas outperformed a traditional conversational UI on the Creativity Support Index. Participants found that it reduced cognitive load, with lower mental demand and frustration. Qualitative feedback revealed that the visual organization of thoughts and easy iteration encouraged new perspectives and ideas. A follow-up field study (N=10) confirmed these results, showcasing the potential of dynamic, customizable interfaces in improving collaborative writing with AI. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:13:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Amin",
"Rifat Mehreen",
""
],
[
"Kühle",
"Oliver Hans",
""
],
[
"Buschek",
"Daniel",
""
],
[
"Butz",
"Andreas",
""
]
] |
2506.03742 | Hao Chen | Hao Chen, Hanyu Wang, Tingting Wang, Yongsen Tang, Haoshu Li, Xiaohong Yan, Lifa Zhang | Ultrafast switching of photoinduced phonon chirality in the antiferrochiral BPO$_{4}$ crystal | 14 pages, 4 figures | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In crystalline systems, chiral crystals cannot interconvert to their enantiomorph post-synthesis without undergoing melting-recrystallization processes. However, recent work indicates that ultrafast terahertz-polarized light has been shown to enable dynamic control of structural chirality in the antiferrochiral boron phosphate (BPO$_4$) crystal. Here, using first-principles calculations and nonlinear phonon dynamics simulations, we investigate the underlying physics of lattice dynamics in this system. The results demonstrate that polarized optical pumping not only induces chiral phonons but also establishes a chirality-selective filtering mechanism, both of which can be reversibly switched by tuning the polarization of the excitation pulse. Furthermore, under a temperature gradient, the pump-induced chiral phonons give rise to ultrafast phonon magnetization, with its direction also controllable via light polarization. Our findings establish a new paradigm for ultrafast optical control of phonon chirality via dynamic chirality switching, offering promising opportunities for chiral information transfer and the design of chiral phononic devices. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:14:05 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Hao",
""
],
[
"Wang",
"Hanyu",
""
],
[
"Wang",
"Tingting",
""
],
[
"Tang",
"Yongsen",
""
],
[
"Li",
"Haoshu",
""
],
[
"Yan",
"Xiaohong",
""
],
[
"Zhang",
"Lifa",
""
]
] |
2506.03743 | Francisco Munguia-Galeano | Francisco Munguia-Galeano, Louis Longley, Satheeshkumar Veeramani, Zhengxue Zhou, Rob Clowes, Hatem Fakhruldeen, and Andrew I. Cooper | An Open-source Capping Machine Suitable for Confined Spaces | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the context of self-driving laboratories (SDLs), ensuring automated and error-free capping is crucial, as it is a ubiquitous step in sample preparation. Automated capping in SDLs can occur in both large and small workspaces (e.g., inside a fume hood). However, most commercial capping machines are designed primarily for large spaces and are often too bulky for confined environments. Moreover, many commercial products are closed-source, which can make their integration into fully autonomous workflows difficult. This paper introduces an open-source capping machine suitable for compact spaces, which also integrates a vision system that recognises capping failure. The capping and uncapping processes are repeated 100 times each to validate the machine's design and performance. As a result, the capping machine reached a 100 % success rate for capping and uncapping. Furthermore, the machine sealing capacities are evaluated by capping 12 vials filled with solvents of different vapour pressures: water, ethanol and acetone. The vials are then weighed every 3 hours for three days. The machine's performance is benchmarked against an industrial capping machine (a Chemspeed station) and manual capping. The vials capped with the prototype lost 0.54 % of their content weight on average per day, while the ones capped with the Chemspeed and manually lost 0.0078 % and 0.013 %, respectively. The results show that the capping machine is a reasonable alternative to industrial and manual capping, especially when space and budget are limitations in SDLs. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:14:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Munguia-Galeano",
"Francisco",
""
],
[
"Longley",
"Louis",
""
],
[
"Veeramani",
"Satheeshkumar",
""
],
[
"Zhou",
"Zhengxue",
""
],
[
"Clowes",
"Rob",
""
],
[
"Fakhruldeen",
"Hatem",
""
],
[
"Cooper",
"Andrew I.",
""
]
] |
2506.03744 | Sebastian Lerch | Tilmann Gneiting, Tobias Biegert, Kristof Kraus, Eva-Maria Walz, Alexander I. Jordan, Sebastian Lerch | Probabilistic measures afford fair comparisons of AIWP and NWP model output | null | null | null | null | stat.AP physics.ao-ph stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a new measure for fair and meaningful comparisons of single-valued output from artificial intelligence based weather prediction (AIWP) and numerical weather prediction (NWP) models, called potential continuous ranked probability score (PC). In a nutshell, we subject the deterministic backbone of physics-based and data-driven models post hoc to the same statistical postprocessing technique, namely, isotonic distributional regression (IDR). Then we find PC as the mean continuous ranked probability score (CRPS) of the postprocessed probabilistic forecasts. The nonnegative PC measure quantifies potential predictive performance and is invariant under strictly increasing transformations of the model output. PC attains its most desirable value of zero if, and only if, the weather outcome Y is a fixed, non-decreasing function of the model output X. The PC measure is recorded in the unit of the outcome, has an upper bound of one half times the mean absolute difference between outcomes, and serves as a proxy for the mean CRPS of real-time, operational probabilistic products. When applied to WeatherBench 2 data, our approach demonstrates that the data-driven GraphCast model outperforms the leading, physics-based European Centre for Medium Range Weather Forecasts (ECMWF) high-resolution (HRES) model. Furthermore, the PC measure for the HRES model aligns exceptionally well with the mean CRPS of the operational ECMWF ensemble. Across application domains, our approach affords comparisons of single-valued forecasts in settings where the pre-specification of a loss function -- which is the usual, and principally superior, procedure in forecast contests, administrative, and benchmarks settings -- places competitors on unequal footings. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:14:45 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gneiting",
"Tilmann",
""
],
[
"Biegert",
"Tobias",
""
],
[
"Kraus",
"Kristof",
""
],
[
"Walz",
"Eva-Maria",
""
],
[
"Jordan",
"Alexander I.",
""
],
[
"Lerch",
"Sebastian",
""
]
] |
2506.03745 | Jules Chenal | Jules Chenal (LPP), Matilde Manzaroli | Real Toric Varieties: Interactions between their Geometry and their Topology | null | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the present article, we investigate the topology of real toric varieties, especially those whose torus is not split over the field of real numbers. We describe some canonical fibrations associated to their real loci. Then, we establish various properties of their cohomology provided that their real loci are compact and smooth. For instance, we compute their Betti numbers, show that their cohomology is totally algebraic, and extend a criterion of orientability. In addition, we provide the topological classification of equivariant embeddings of non-split tridimensional tori. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:15:35 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chenal",
"Jules",
"",
"LPP"
],
[
"Manzaroli",
"Matilde",
""
]
] |
2506.03746 | C\'esar Sabater | C\'esar Sabater, Sonia Ben Mokhtar, Jan Ramon | Dropout-Robust Mechanisms for Differentially Private and Fully Decentralized Mean Estimation | 23 pages, 4 figures | null | null | null | cs.CR cs.DC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Achieving differentially private computations in decentralized settings poses significant challenges, particularly regarding accuracy, communication cost, and robustness against information leakage. While cryptographic solutions offer promise, they often suffer from high communication overhead or require centralization in the presence of network failures. Conversely, existing fully decentralized approaches typically rely on relaxed adversarial models or pairwise noise cancellation, the latter suffering from substantial accuracy degradation if parties unexpectedly disconnect. In this work, we propose IncA, a new protocol for fully decentralized mean estimation, a widely used primitive in data-intensive processing. Our protocol, which enforces differential privacy, requires no central orchestration and employs low-variance correlated noise, achieved by incrementally injecting sensitive information into the computation. First, we theoretically demonstrate that, when no parties permanently disconnect, our protocol achieves accuracy comparable to that of a centralized setting-already an improvement over most existing decentralized differentially private techniques. Second, we empirically show that our use of low-variance correlated noise significantly mitigates the accuracy loss experienced by existing techniques in the presence of dropouts. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:16:34 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sabater",
"César",
""
],
[
"Mokhtar",
"Sonia Ben",
""
],
[
"Ramon",
"Jan",
""
]
] |
2506.03747 | YingJie Shi | Yingjie Shi, Jinye Miao, Taotao Qin, Fuyao Cai, Yi Wei, Lingfeng Liu, Tongyao Li, Chenyang Wu, Huan Liang, Yuyang Yin, Lianfa Bai, Enlai Guo, Jing Han | Fast Non-Line-of-Sight Transient Data Simulation and an Open Benchmark Dataset | null | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | Non-Line-of-Sight (NLOS) imaging reconstructs the shape and depth of hidden objects from picosecond-resolved transient signals, offering potential applications in autonomous driving, security, and medical diagnostics. However, current NLOS experiments rely on expensive hardware and complex system alignment, limiting their scalability. This manuscript presents a simplified simulation method that generates NLOS transient data by modeling light-intensity transport rather than performing conventional path tracing, significantly enhancing computational efficiency. All scene elements, including the relay surface, hidden target, stand-off distance, detector time resolution, and acquisition window are fully parameterized, allowing for rapid configuration of test scenarios. Reconstructions based on the simulated data accurately recover hidden geometries, validating the effectiveness of the approach. The proposed tool reduces the entry barrier for NLOS research and supports the optimization of system design. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:17:12 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shi",
"Yingjie",
""
],
[
"Miao",
"Jinye",
""
],
[
"Qin",
"Taotao",
""
],
[
"Cai",
"Fuyao",
""
],
[
"Wei",
"Yi",
""
],
[
"Liu",
"Lingfeng",
""
],
[
"Li",
"Tongyao",
""
],
[
"Wu",
"Chenyang",
""
],
[
"Liang",
"Huan",
""
],
[
"Yin",
"Yuyang",
""
],
[
"Bai",
"Lianfa",
""
],
[
"Guo",
"Enlai",
""
],
[
"Han",
"Jing",
""
]
] |
2506.03748 | Hao Chen | Hao Chen, Haoshu Li, Lifa Zhang, Qian Niu | Magic of nonlocal geometric force: lighting up optical transition and transporting angular momentum by chiral phonons | 6 pages, 4 figures | null | null | null | cond-mat.mtrl-sci cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the impact of the nonlocal geometric force -- arising from the molecular Berry curvature -- on the lattice dynamics of magnetic materials with broken time-reversal symmetry. A first-principles computational framework is established to evaluate this force across the entire Brillouin zone. We apply it to monolayer CoCl$_2$, a ferromagnetic half-semiconductor with a narrow bandgap forbidding direct dipolar optical transition. At the phonon Brillouin zone center, the pronounced nonlocal geometric force leads to a splitting of the two upper optical phonon branches by $3 \times 10^{-2}$ THz, transforming the phonons into chiral modes. Optical chiral phonons can light up the intravalley dark exciton via absorpting circularly polarized photons. Furthermore, acoustic chiral phonons induced by the nonlocal geometric force can transport angular momentum and contribute to a non-dissipative phonon Hall viscosity. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:17:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Hao",
""
],
[
"Li",
"Haoshu",
""
],
[
"Zhang",
"Lifa",
""
],
[
"Niu",
"Qian",
""
]
] |
2506.03749 | Athanase Papadopoulos | Ismail Saglam, Ken'Ichi Ohshika, Athanase Papadopoulos (IRMA) | On families of Finsler metrics | null | null | null | null | math.DG math.GT math.MG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we answer some natural questions on symmetrisation and more general combinations of Finsler metrics, with a view towards applications to Funk and Hilbert geometries and to metrics on Teichm{\"u}ller spaces. For a general non-symmetric Finsler metric on a smooth manifold, we introduce two different families of metrics, containing as special cases the arithmetic and the max symmetrisations respectively of the distance functions associated with these Finsler metrics. We are interested in various natural questions concerning metrics in such a family, regarding its geodesics, its completeness, conditions under which such a metric is Finsler, the shape of its unit ball in the case where it is Finsler, etc. We address such questions in particular in the setting of Funk and Hilbert geometries, and in that of the Teichm{\"u}ller spaces of several kinds of surfaces, equipped with Thurstonlike asymmetric metrics. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:17:46 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Saglam",
"Ismail",
"",
"IRMA"
],
[
"Ohshika",
"Ken'Ichi",
"",
"IRMA"
],
[
"Papadopoulos",
"Athanase",
"",
"IRMA"
]
] |
2506.03750 | Mengxi Xiao | Mengxi Xiao, Mang Ye, Ben Liu, Xiaofen Zong, He Li, Jimin Huang, Qianqian Xie, Min Peng | A Retrieval-Augmented Multi-Agent Framework for Psychiatry Diagnosis | 40 pages, 11 figures | null | null | null | cs.SI cs.AI | http://creativecommons.org/licenses/by/4.0/ | The application of AI in psychiatric diagnosis faces significant challenges, including the subjective nature of mental health assessments, symptom overlap across disorders, and privacy constraints limiting data availability. To address these issues, we present MoodAngels, the first specialized multi-agent framework for mood disorder diagnosis. Our approach combines granular-scale analysis of clinical assessments with a structured verification process, enabling more accurate interpretation of complex psychiatric data. Complementing this framework, we introduce MoodSyn, an open-source dataset of 1,173 synthetic psychiatric cases that preserves clinical validity while ensuring patient privacy. Experimental results demonstrate that MoodAngels outperforms conventional methods, with our baseline agent achieving 12.3% higher accuracy than GPT-4o on real-world cases, and our full multi-agent system delivering further improvements. Evaluation in the MoodSyn dataset demonstrates exceptional fidelity, accurately reproducing both the core statistical patterns and complex relationships present in the original data while maintaining strong utility for machine learning applications. Together, these contributions provide both an advanced diagnostic tool and a critical research resource for computational psychiatry, bridging important gaps in AI-assisted mental health assessment. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:18:25 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xiao",
"Mengxi",
""
],
[
"Ye",
"Mang",
""
],
[
"Liu",
"Ben",
""
],
[
"Zong",
"Xiaofen",
""
],
[
"Li",
"He",
""
],
[
"Huang",
"Jimin",
""
],
[
"Xie",
"Qianqian",
""
],
[
"Peng",
"Min",
""
]
] |
2506.03751 | Sangita Yadav | Ankit Kumar, Sarvesh Kumar, and Sangita Yadav | Convergence Analysis of Virtual Element Methods for the Sobolev Equation with Convection | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We explore the potential applications of virtual elements for solving the Sobolev equation with a convective term. A conforming virtual element method is employed for spatial discretization, while an implicit Euler scheme is used to approximate the time derivative. To establish the optimal rate of convergence, a novel intermediate projection operator is introduced. We discuss and analyze both the semi-discrete and fully discrete schemes, deriving optimal error estimates for both the energy norm and L2-norm. Several numerical experiments are conducted to validate the theoretical findings and assess the computational efficiency of the proposed numerical methods. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:18:35 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kumar",
"Ankit",
""
],
[
"Kumar",
"Sarvesh",
""
],
[
"Yadav",
"Sangita",
""
]
] |
2506.03752 | Goncalo Mesquita | Gon\c{c}alo Mesquita, Ana Rita C\'oias, Artur Dubrawski and Alexandre Bernardino | Frame-Level Real-Time Assessment of Stroke Rehabilitation Exercises from Video-Level Labeled Data: Task-Specific vs. Foundation Models | null | null | null | null | eess.IV | http://creativecommons.org/licenses/by/4.0/ | The growing demands of stroke rehabilitation have increased the need for solutions to support autonomous exercising. Virtual coaches can provide real-time exercise feedback from video data, helping patients improve motor function and keep engagement. However, training real-time motion analysis systems demands frame-level annotations, which are time-consuming and costly to obtain. In this work, we present a framework that learns to classify individual frames from video-level annotations for real-time assessment of compensatory motions in rehabilitation exercises. We use a gradient-based technique and a pseudo-label selection method to create frame-level pseudo-labels for training a frame-level classifier. We leverage pre-trained task-specific models - Action Transformer, SkateFormer - and a foundation model - MOMENT - for pseudo-label generation, aiming to improve generalization to new patients. To validate the approach, we use the \textit{SERE} dataset with 18 post-stroke patients performing five rehabilitation exercises annotated on compensatory motions. MOMENT achieves better video-level assessment results (AUC = $73\%$), outperforming the baseline LSTM (AUC = $58\%$). The Action Transformer, with the Integrated Gradient technique, leads to better outcomes (AUC = $72\%$) for frame-level assessment, outperforming the baseline trained with ground truth frame-level labeling (AUC = $69\%$). We show that our proposed approach with pre-trained models enhances model generalization ability and facilitates the customization to new patients, reducing the demands of data labeling. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:20:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mesquita",
"Gonçalo",
""
],
[
"Cóias",
"Ana Rita",
""
],
[
"Dubrawski",
"Artur",
""
],
[
"Bernardino",
"Alexandre",
""
]
] |
2506.03754 | Alexander V. Karzanov | Vladimir I. Danilov, Alexander V. Karzanov, Gleb A. Koshevoy | On universal quadratic inequalities for minors of TNN matrices | 10 pages, 5 figures | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this note, for positive integers $n,n'$, we give a geometric-combinatorial characterization for the set of quadratic inequalities on minors that are valid for all $n\times n'$ totally nonnegative matrices. This is obtained as a consequence from our earlier results on ``stable'' quadratic identities on minors of matrices generated by flows in planar graphs via Lindstr\"om's construction. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:22:26 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Danilov",
"Vladimir I.",
""
],
[
"Karzanov",
"Alexander V.",
""
],
[
"Koshevoy",
"Gleb A.",
""
]
] |
2506.03755 | Max Hellrigel-Holderbaum | Max Hellrigel-Holderbaum and Leonard Dung | Misalignment or misuse? The AGI alignment tradeoff | Forthcoming in Philosophical Studies | null | null | null | cs.CY cs.AI | http://creativecommons.org/licenses/by/4.0/ | Creating systems that are aligned with our goals is seen as a leading approach to create safe and beneficial AI in both leading AI companies and the academic field of AI safety. We defend the view that misaligned AGI - future, generally intelligent (robotic) AI agents - poses catastrophic risks. At the same time, we support the view that aligned AGI creates a substantial risk of catastrophic misuse by humans. While both risks are severe and stand in tension with one another, we show that - in principle - there is room for alignment approaches which do not increase misuse risk. We then investigate how the tradeoff between misalignment and misuse looks empirically for different technical approaches to AI alignment. Here, we argue that many current alignment techniques and foreseeable improvements thereof plausibly increase risks of catastrophic misuse. Since the impacts of AI depend on the social context, we close by discussing important social factors and suggest that to reduce the risk of a misuse catastrophe due to aligned AGI, techniques such as robustness, AI control methods and especially good governance seem essential. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:22:37 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Hellrigel-Holderbaum",
"Max",
""
],
[
"Dung",
"Leonard",
""
]
] |
2506.03756 | Romanshu Garg | Shivank Pal, Romanshu Garg, Gyan Prakash Singh, Gauree Shanker | Cosmological Dynamics of Accelerating Model in $f(T)$ Gravity with Special Form of Deceleration Parameter | 18 pages, 6 figures | null | null | null | gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, the dynamical behavior of the accelerated expansion of the universe is discussed within the framework of $f(T)$ gravity for a particular form of $ f(T)=\alpha (-T)^{n}$. Considering a redshift-dependent parameterization of the deceleration parameter as $q(z)=q_{0}+ q_{1}\left(\frac{ \ln(N+z)}{z+1}-\ln N \right)$. We have derived the Hubble parameter in terms of redshift and discussed its effect on other cosmological parameters. Using Bayesian statistical analysis and $\chi^2$-minimization, the median values of the model parameters for both the cosmic chronometer and the combined (CC + Pantheon) dataset have been determined. Further, energy density, pressure, the equation of state for Dark Energy and statefinder diagnostics are analyzed. The current age of the universe is also computed for this model. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:22:54 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pal",
"Shivank",
""
],
[
"Garg",
"Romanshu",
""
],
[
"Singh",
"Gyan Prakash",
""
],
[
"Shanker",
"Gauree",
""
]
] |
2506.03757 | Razvan-Andrei Lascu | Razvan-Andrei Lascu, David \v{S}i\v{s}ka, {\L}ukasz Szpruch | PPO in the Fisher-Rao geometry | 17 pages | null | null | null | cs.LG math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Proximal Policy Optimization (PPO) has become a widely adopted algorithm for reinforcement learning, offering a practical policy gradient method with strong empirical performance. Despite its popularity, PPO lacks formal theoretical guarantees for policy improvement and convergence. PPO is motivated by Trust Region Policy Optimization (TRPO) that utilizes a surrogate loss with a KL divergence penalty, which arises from linearizing the value function within a flat geometric space. In this paper, we derive a tighter surrogate in the Fisher-Rao (FR) geometry, yielding a novel variant, Fisher-Rao PPO (FR-PPO). Our proposed scheme provides strong theoretical guarantees, including monotonic policy improvement. Furthermore, in the tabular setting, we demonstrate that FR-PPO achieves sub-linear convergence without any dependence on the dimensionality of the action or state spaces, marking a significant step toward establishing formal convergence results for PPO-based algorithms. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:23:27 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Lascu",
"Razvan-Andrei",
""
],
[
"Šiška",
"David",
""
],
[
"Szpruch",
"Łukasz",
""
]
] |
2506.03758 | Daniel Palenicek | Daniel Palenicek, Florian Vogt, Jan Peters | Scaling CrossQ with Weight Normalization | arXiv admin note: substantial text overlap with arXiv:2502.07523 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Reinforcement learning has achieved significant milestones, but sample efficiency remains a bottleneck for real-world applications. Recently, CrossQ has demonstrated state-of-the-art sample efficiency with a low update-to-data (UTD) ratio of 1. In this work, we explore CrossQ's scaling behavior with higher UTD ratios. We identify challenges in the training dynamics which are emphasized by higher UTDs, particularly Q-bias explosion and the growing magnitude of critic network weights. To address this, we integrate weight normalization into the CrossQ framework, a solution that stabilizes training, prevents potential loss of plasticity and keeps the effective learning rate constant. Our proposed approach reliably scales with increasing UTD ratios, achieving competitive or superior performance across a range of challenging tasks on the DeepMind control benchmark, notably the complex dog and humanoid environments. This work eliminates the need for drastic interventions, such as network resets, and offers a robust pathway for improving sample efficiency and scalability in model-free reinforcement learning. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:24:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Palenicek",
"Daniel",
""
],
[
"Vogt",
"Florian",
""
],
[
"Peters",
"Jan",
""
]
] |
2506.03759 | Matteo Della Rossa | Thiago Alves Lima and Matteo Della Rossa and Antoine Girard | Feedback stabilization of switched systems under arbitrary switching: A convex characterization | Submitted to Automatica | null | null | null | math.OC cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study stabilizability of discrete-time switched linear systems where the switching signal is considered as an arbitrary disturbance (and not a control variable). We characterize feedback stabilization via necessary and sufficient linear matrix inequalities (LMIs) conditions based on novel graph structures. We analyze both the cases in which the controller has (or has not) access to the current switching mode, the so-called mode-dependent and mode-independent settings, providing specular results. Moreover, our approach provides explicit piecewise-linear and memory-dependent linear controllers, highlighting the connections with existing stabilization approaches. The effectiveness of the proposed technique is finally illustrated with the help of some numerical examples. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:24:59 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Lima",
"Thiago Alves",
""
],
[
"Della Rossa",
"Matteo",
""
],
[
"Girard",
"Antoine",
""
]
] |
2506.03760 | Changmin Park | Changmin Park, Beomjoon Lee, Haechan Jung, Haejin Jung, Changjoo Nam | Understanding Physical Properties of Unseen Deformable Objects by Leveraging Large Language Models and Robot Actions | null | null | null | null | cs.RO cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we consider the problem of understanding the physical properties of unseen objects through interactions between the objects and a robot. Handling unseen objects with special properties such as deformability is challenging for traditional task and motion planning approaches as they are often with the closed world assumption. Recent results in Large Language Models (LLMs) based task planning have shown the ability to reason about unseen objects. However, most studies assume rigid objects, overlooking their physical properties. We propose an LLM-based method for probing the physical properties of unseen deformable objects for the purpose of task planning. For a given set of object properties (e.g., foldability, bendability), our method uses robot actions to determine the properties by interacting with the objects. Based on the properties examined by the LLM and robot actions, the LLM generates a task plan for a specific domain such as object packing. In the experiment, we show that the proposed method can identify properties of deformable objects, which are further used for a bin-packing task where the properties take crucial roles to succeed. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:25:12 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Park",
"Changmin",
""
],
[
"Lee",
"Beomjoon",
""
],
[
"Jung",
"Haechan",
""
],
[
"Jung",
"Haejin",
""
],
[
"Nam",
"Changjoo",
""
]
] |
2506.03761 | Hongcheng Guo | Hongcheng Guo, Zheyong Xie, Shaosheng Cao, Boyang Wang, Weiting Liu, Zheyu Ye, Zhoujun Li, Zuozhu Liu | Act-as-Pet: Benchmarking the Abilities of Large Language Models as E-Pets in Social Network Services | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As interest in using Large Language Models (LLMs) for interactive and emotionally rich experiences grows, virtual pet companionship emerges as a novel yet underexplored application. Existing approaches focus on basic pet role-playing interactions without systematically benchmarking LLMs for comprehensive companionship. In this paper, we introduce Pet-Bench, a dedicated benchmark that evaluates LLMs across both self-interaction and human-interaction dimensions. Unlike prior work, Pet-Bench emphasizes self-evolution and developmental behaviors alongside interactive engagement, offering a more realistic reflection of pet companionship. It features diverse tasks such as intelligent scheduling, memory-based dialogues, and psychological conversations, with over 7,500 interaction instances designed to simulate complex pet behaviors. Evaluation of 28 LLMs reveals significant performance variations linked to model size and inherent capabilities, underscoring the need for specialized optimization in this domain. Pet-Bench serves as a foundational resource for benchmarking pet-related LLM abilities and advancing emotionally immersive human-pet interactions. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:25:52 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Guo",
"Hongcheng",
""
],
[
"Xie",
"Zheyong",
""
],
[
"Cao",
"Shaosheng",
""
],
[
"Wang",
"Boyang",
""
],
[
"Liu",
"Weiting",
""
],
[
"Ye",
"Zheyu",
""
],
[
"Li",
"Zhoujun",
""
],
[
"Liu",
"Zuozhu",
""
]
] |
2506.03762 | Yifeng Gu | Yifeng Gu, Zicong Jiang, Jianxiu Jin, Kailing Guo, Ziyang Zhang, Xiangmin Xu | AhaKV: Adaptive Holistic Attention-Driven KV Cache Eviction for Efficient Inference of Large Language Models | 14 pages, 8 figures | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have significantly advanced the field of Artificial Intelligence. However, their deployment is resource-intensive, not only due to the large number of model parameters but also because the (Key-Value) KV cache consumes a lot of memory during inference. While several works propose reducing the KV cache by evicting the unnecessary tokens, these approaches rely on accumulated attention score as eviction score to quantify the importance of the token. We identify the accumulated attention score is biased and it decreases with the position of the tokens in the mathematical expectation. As a result, the retained tokens concentrate on the initial positions, limiting model's access to global contextual information. To address this issue, we propose Adaptive holistic attention KV (AhaKV), it addresses the bias of the accumulated attention score by adaptively tuning the scale of softmax according the expectation of information entropy of attention scores. To make use of the holistic attention information in self-attention mechanism, AhaKV utilize the information of value vectors, which is overlooked in previous works, to refine the adaptive score. We show theoretically that our method is well suited for bias reduction. We deployed AhaKV on different models with a fixed cache budget. Experiments show that AhaKV successfully mitigates bias and retains crucial tokens across global context and achieve state-of-the-art results against other related work on several benchmark tasks. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:25:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gu",
"Yifeng",
""
],
[
"Jiang",
"Zicong",
""
],
[
"Jin",
"Jianxiu",
""
],
[
"Guo",
"Kailing",
""
],
[
"Zhang",
"Ziyang",
""
],
[
"Xu",
"Xiangmin",
""
]
] |
2506.03763 | Quang Hieu Pham | Quang Hieu Pham, Thuy Duong Nguyen, Tung Pham, Anh Tuan Luu, Dat Quoc Nguyen | ClozeMath: Improving Mathematical Reasoning in Language Models by Learning to Fill Equations | Accepted to ACL 2025 Findings | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | The capabilities of large language models (LLMs) have been enhanced by training on data that reflects human thought processes, such as the Chain-of-Thought format. However, evidence suggests that the conventional scheme of next-word prediction may not fully capture how humans learn to think. Inspired by how humans generalize mathematical reasoning, we propose a new approach named ClozeMath to fine-tune LLMs for mathematical reasoning. Our ClozeMath involves a text-infilling task that predicts masked equations from a given solution, analogous to cloze exercises used in human learning. Experiments on GSM8K, MATH, and GSM-Symbolic show that ClozeMath surpasses the strong baseline Masked Thought in performance and robustness, with two test-time scaling decoding algorithms, Beam Search and Chain-of-Thought decoding. Additionally, we conduct an ablation study to analyze the effects of various architectural and implementation choices on our approach. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:27:21 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pham",
"Quang Hieu",
""
],
[
"Nguyen",
"Thuy Duong",
""
],
[
"Pham",
"Tung",
""
],
[
"Luu",
"Anh Tuan",
""
],
[
"Nguyen",
"Dat Quoc",
""
]
] |
2506.03765 | Sicong Han | Sicong Han, Chenhao Lin, Zhengyu Zhao, Xiyuan Wang, Xinlei He, Qian Li, Cong Wang, Qian Wang, Chao Shen | Prediction Inconsistency Helps Achieve Generalizable Detection of Adversarial Examples | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Adversarial detection protects models from adversarial attacks by refusing suspicious test samples. However, current detection methods often suffer from weak generalization: their effectiveness tends to degrade significantly when applied to adversarially trained models rather than naturally trained ones, and they generally struggle to achieve consistent effectiveness across both white-box and black-box attack settings. In this work, we observe that an auxiliary model, differing from the primary model in training strategy or model architecture, tends to assign low confidence to the primary model's predictions on adversarial examples (AEs), while preserving high confidence on normal examples (NEs). Based on this discovery, we propose Prediction Inconsistency Detector (PID), a lightweight and generalizable detection framework to distinguish AEs from NEs by capturing the prediction inconsistency between the primal and auxiliary models. PID is compatible with both naturally and adversarially trained primal models and outperforms four detection methods across 3 white-box, 3 black-box, and 1 mixed adversarial attacks. Specifically, PID achieves average AUC scores of 99.29\% and 99.30\% on CIFAR-10 when the primal model is naturally and adversarially trained, respectively, and 98.31% and 96.81% on ImageNet under the same conditions, outperforming existing SOTAs by 4.70%$\sim$25.46%. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:29:11 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Han",
"Sicong",
""
],
[
"Lin",
"Chenhao",
""
],
[
"Zhao",
"Zhengyu",
""
],
[
"Wang",
"Xiyuan",
""
],
[
"He",
"Xinlei",
""
],
[
"Li",
"Qian",
""
],
[
"Wang",
"Cong",
""
],
[
"Wang",
"Qian",
""
],
[
"Shen",
"Chao",
""
]
] |
2506.03766 | Vicente Jos\'e Bol\'os | Vicente J. Bolos, Rafael Benitez, Vicente Coll-Serrano | Conventional and Fuzzy Data Envelopment Analysis with deaR | 41 pages, 9 figures | null | null | null | econ.EM cs.MS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | deaR is a recently developed R package for data envelopment analysis (DEA) that implements a large number of conventional and fuzzy models, along with super-efficiency models, cross-efficiency analysis, Malmquist index, bootstrapping, and metafrontier analysis. It should be noted that deaR is the only package to date that incorporates Kao-Liu, Guo-Tanaka and possibilistic fuzzy models. The versatility of the package allows the user to work with different returns to scale and orientations, as well as to consider special features, namely non-controllable, non-discretionary or undesirable variables. Moreover, it includes novel graphical representations that can help the user to display the results. This paper is a comprehensive description of deaR, reviewing all implemented models and giving examples of use. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:29:23 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bolos",
"Vicente J.",
""
],
[
"Benitez",
"Rafael",
""
],
[
"Coll-Serrano",
"Vicente",
""
]
] |
2506.03767 | Loic Foissy | Lo\"ic Foissy (LMPA) | Construction of pre-and post-Lie algebras for stochastic PDEs | null | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We give and study a construction of pre-Lie algebra structures on rooted trees whose edges and vertices are decorated, with a grafting product acting, through a map $\phi$, both on the decoration of the created edge and on the vertex that holds the grafting. We show that this construction gives a pre-Lie algebra if, and only if, the map $\phi$ satisfies a commutation relation, called tree-compatibility. We show how to extend this pre-Lie algebra structure to a post-Lie one by a semi-direct extension with another post-Lie algebra. We also define several constructions to obtain tree-compatible maps, and give examples, including a description of all tree-admissible maps when the space of decorations of the vertices is $2$-dimensional and the space of decorations of the edges is finite-dimensional. A particular example of such a construction is used by Bruned, Hairer and Zambotti for the study of stochastic partial differential equations: when no noise is involved, we show that the underlying tree-compatible map is the exponential of a simpler one and deduce an explicit isomorphism with a classical pre-Lie algebra of rooted trees; when a noise is involved, we obtain the underlying tree-compatible map as a direct sum. We also obtain with our formalism the post-Lie algebras described by Bruned and Katsetsiadis. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:30:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Foissy",
"Loïc",
"",
"LMPA"
]
] |
2506.03768 | Matt Sampson | Matt L. Sampson, James R. Beattie, Romain Teyssier, Philipp Kempski, Eric R. Moseley, Beno\^it Commer\c{c}on, Yohan Dubois, and Joakim Rosdahl | Cosmic ray and plasma coupling for isothermal supersonic turbulence in the magnetized interstellar medium | 26 pages, 11 figures, submitted to ApJ | null | null | null | astro-ph.GA astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Cosmic rays (CRs) are an integral part of the non-thermal pressure budget in the interstellar medium (ISM) and are the leading-order ionization mechanism in cold molecular clouds. We study the impacts that different microphysical CR diffusion coefficients and streaming speeds have on the evolution of isothermal, magnetized, turbulent plasmas, relevant to the cold ISM. We utilized a two-moment CR magnetohydrodynamic (CRMHD) model, allowing us to dynamically evolve both CR energy and flux densities with contributions from Alfv\'enic streaming and anisotropic diffusion. We identify $\textit{coupled}$ and $\textit{decoupled}$ regimes, and define dimensionless Prandtl numbers $\rm{Pm_c}$ and $\rm{Pm_s}$, which quantify whether the plasma falls within these two regimes. In the coupled regime -- characteristic of slow streaming ($\rm{Pm_s} < 1$) and low diffusion ($\rm{Pm_c} < 1$) -- the CR fluid imprints upon the plasma a mixed equation of state between $P_{\rm{c}} \propto \rho^{4/3}$ (relativistic fluid) and $P_{\rm{c}} \propto \rho^{2/3}$ (streaming), where $P_{\rm{c}}$ is the CR pressure, and $\rho$ is the plasma density. By modifying the sound speed, the coupling reduces the turbulent Mach number, and hence the amplitude of the density fluctuations, whilst supporting secular heating of the CR fluid. In contrast, in the decoupled regime ($\rm{Pm_s} > 1$ or $\rm{Pm_c} > 1$) the CR fluid and the plasma have negligible interactions. We further show that CR heating is enabled by coherent structures within the compressible velocity field, with no impact on the turbulence spectrum of incompressible modes. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:30:19 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sampson",
"Matt L.",
""
],
[
"Beattie",
"James R.",
""
],
[
"Teyssier",
"Romain",
""
],
[
"Kempski",
"Philipp",
""
],
[
"Moseley",
"Eric R.",
""
],
[
"Commerçon",
"Benoît",
""
],
[
"Dubois",
"Yohan",
""
],
[
"Rosdahl",
"Joakim",
""
]
] |
2506.03769 | Laurent Pedesseau | Sreejith Pallikkara Chandrasekharan, Sofia Apergi, Charles Cornet, Laurent Pedesseau | Efficient and standardized interface energy calculations in hybrid heterostructures using fictitious atoms surface passivation | 24 pages, 3 figures | null | null | null | cond-mat.mtrl-sci physics.app-ph physics.chem-ph physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | Heterostructures combining diverse physico-chemical properties are increasingly in demand for a wide range of applications in modern science and technology. However, despite their importance in materials science, accurately determining absolute interface energies remains a major challenge. This difficulty arises from periodic boundary conditions, high computational costs of plane-wave methods, multipolar interactions in heterostructures, the need for thick slabs for interface convergence, and reconstructed surfaces on both slab faces. Here, we introduce a standardized and computationally efficient fictitious H* charge passivation method for the surface termination, designed to accurately determine absolute interface energies in heterogeneous materials associations. This approach effectively addresses issues associated with surface reconstructions while significantly reducing computational costs within the framework of density functional theory. To demonstrate its reliability, we calculate the absolute interface energies for various quasi-lattice-matched and lattice-mismatched abrupt III-V/Si interfaces using the H* passivation technique and benchmark the results against those obtained using conventional reconstructed surface methods. We further explore the early stages of strained epitaxial GaAs on Si(001). Finally, we assess the fictitious H* passivation method, showing its effectiveness in minimizing electric dipole errors, reducing computational costs, and thus decreasing greenhouse gas emissions from high-performance computing. Finally, the potential of the approach to compute interface energies across a broad spectrum of materials is emphasized. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:30:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chandrasekharan",
"Sreejith Pallikkara",
""
],
[
"Apergi",
"Sofia",
""
],
[
"Cornet",
"Charles",
""
],
[
"Pedesseau",
"Laurent",
""
]
] |
2506.03770 | Mingjun Sun | Mingjun Sun, Chongjun Ouyang, Shaochuan Wu, and Yuanwei Liu | Multiuser Beamformig for Pinching-Antenna Systems: An Element-wise Optimization Framework | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The pinching-antenna system (PASS) reconstructs wireless channels through pinching beamforming, i.e., optimizing the activated locations of pinching antennas (PAs) along the waveguide. The aim of this article is to investigate the joint design of baseband beamforming and pinching beamforming. A low-complexity element-wise sequential optimization framework is proposed to address the sum-rate maximization problem in PASS-enabled downlink and uplink channels. i) For the downlink scenario, maximum ratio transmission (MRT), zero-forcing (ZF), and minimum mean square error (MMSE) beamforming schemes are employed as baseband beamformers. For each beamformer, a closed-form expression for the downlink sum-rate is derived as a single-variable function with respect to the pinching beamformer. Based on this, a sequential optimization method is proposed, where the positions of the PAs are updated element-wise using a low-complexity one-dimensional search. ii) For the uplink scenario, signal detection is performed using maximum ratio combining (MRC), ZF, and MMSE combiners. A closed-form sum-rate expression is derived for each linear combiner, and a similar element-wise design is applied to optimize the pinching beamforming. Numerical results are provided to validate the effectiveness of the proposed method and demonstrate that: (i) For all considered linear beamformers, the proposed PASS architecture outperforms conventional fixed-antenna systems in terms of sum-rate performance; (ii) in both downlink and uplink channels, ZF achieves performance close to that of MMSE and significantly outperforms MRT or MRC; and (iii) the proposed element-wise design eliminates the need for alternating updates between the baseband and pinching beamformers, thereby ensuring low computational complexity. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:30:58 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Sun",
"Mingjun",
""
],
[
"Ouyang",
"Chongjun",
""
],
[
"Wu",
"Shaochuan",
""
],
[
"Liu",
"Yuanwei",
""
]
] |
2506.03771 | Tatpong Katanyukul | Tatpong Katanyukul | Toward Entailment Checking: Explore Eigenmarking Search | 8 pages | null | null | null | quant-ph cs.ET cs.LO | http://creativecommons.org/licenses/by/4.0/ | Logic entailment is essential to reasoning,
but entailment checking has the worst-case complexity of an exponential of the variable size.
With recent development, quantum computing when mature
may allow an effective approach for various combinatorial problems,
including entailment checking.
Grover algorithm uses Grover operations, selective phase inversion and amplitude amplification
to address a search over unstructured data with quadratic improvement from a classical method.
Its original form is intended to a single-winner scenario: exactly one match is promised.
Its extension to multiple-winner cases employs probabilistic control over
a number of applications of Grover operations, while a no-winner case is handled by time-out.
Our study explores various schemes of ``eigenmarking'' approach.
Still relying on Grover operations, but the approach introduces additional qubits
to tag the eigenstates.
The tagged eigenstates are to facilitate an interpretation of the measured results and
enhance identification of a no-winner case (related to no logic violation in entailment context).
Our investigation experiments three variations of eigenmarking on a two-qubit system
using an IBM Aer simulator.
The results show strong distinguishability in all schemes with the best relative distinguishabilities of 19 and 53 in worst case
and in average case, respectively.
Our findings reveal a viable quantum mechanism to differentiate a no-winner case
from other scenarios, which could play a pivot role in entailment checking and logic reasoning in general. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:33:10 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Katanyukul",
"Tatpong",
""
]
] |
2506.03772 | Swagata Ghosh | Swagata Ghosh | Fermionic decay of light charged pseudoscalar in Georgi Machacek model | Contribution to : PPC 2024 | null | 10.1007/978-981-96-4986-0_37 | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | The Georgi Machacek (GM) model is a triplet extension of the scalar sector of the Standard Model (SM) containing charged and neutral scalars and pseudoscalars with alluring phenomenology. The CMS and ATLAS collaborations of the LHC at $\sqrt{s}=13$ TeV already searched for various decays of the charged Higgs boson in low as well as high mass region. The low mass charged Higgs are produced by the decay of top quarks or antiquarks. $H^{\pm}\rightarrow \tau\nu,\, cs$ are the main fermionic decay channels of light charged scalars or pseudoscalars. For low mass region of the charged pseudoscalars, the triplet vev is restricted from above from the indirect search coming from the $b\rightarrow s \gamma$ decay. Here, I consider the observed data from the ATLAS and CMS both for the light charged Higgs decaying into $cs$ and $\tau\nu$. The ATLAS data observed for $H^{\pm}\rightarrow \tau\nu$ offers a more stringent bound on $v_2$ for GM model pseudoscalar mass below $160$ GeV. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:35:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ghosh",
"Swagata",
""
]
] |
2506.03773 | Mohsen Kian | Mohsen Kian | Deviation from Complete Positivity: Structural Insights and Quantum Information Applications | null | null | null | null | quant-ph math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce the CP-distance as a measure of how far a Hermitian map is from being completely positive, deriving key properties and bounds. We investigate the role of CP-distance in the structural analysis of positive Hermitian linear maps between matrix algebras, focusing on its implications for quantum information theory. In particular, we derive bounds on the detection strength of entanglement witnesses. We elucidate the interplay between CP-distance and the structural properties of positive maps, offering insights into their decompositions. We also analyze how the CP-distance influences the decompositions of positive Hermitian maps, revealing its impact on the balance between completely positive components. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:36:41 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kian",
"Mohsen",
""
]
] |
2506.03774 | Santabrata Das | Seshadri Majumder (IITG), Ankur Kushwaha (URSC), Swapnil Singh (URSC), Kiran M. Jayasurya (URSC), Santabrata Das (IITG), Anuj Nandi (URSC) | Probing the accretion geometry of black hole X-ray binaries: A multi-mission spectro-polarimetric and timing study | 26 pages, 8 figures, 6 tables, Comments welcome | null | null | null | astro-ph.HE astro-ph.IM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a comprehensive spectro-polarimetric and timing analysis of eleven black hole X-ray binaries, namely Cyg X-1, 4U 1630-47, Cyg X-3, LMC X-1, 4U 1957+115, LMC X-3, Swift J1727.8-1613, GX 339-4, Swift J151857.0-572147, IGR J17091-3624, and MAXI J1744-294, using quasi-simultaneous observations from {\it IXPE}, {\it NICER}, {\it NuSTAR}, and {\it AstroSat}. Timing analyses reveal characteristic type-B and type-C Quasi-periodic Oscillations (QPOs) across different spectral states, often associated with episodic radio ejections. Broadband ($0.7-60$ keV) spectral modelling, employing disc, Comptonization, and reflection components, reveals degeneracies in constraining disc-corona geometries. Polarimetric measurements in 2-8 keV band detect significant polarization degrees (PDs) ranging from $3-20.6\%$ ($1.2-21.4\%$) in harder (softer) states, with moderate to strong energy dependence, except for LMC X-1, Swift J151857.0-572147, and MAXI J1744-294, where no significant polarization is detected. We report the polarization detections of Cyg X-3 (PD $\sim 21.4\%$, SIMS), LMC X-3 (PD $\sim 2.4\%$, HSS) and IGR J17091-3624 (PD $\sim 9\%$, LHS) using the recent {\it IXPE} observations. A positive correlation is found between PD and the Comptonized photon fraction ($cov_{\rm frac}$), while an anti-correlation is observed with the disc-to-Comptonized flux ratio ($F_{\rm ratio}$) across spectral states. The combined timing, spectral, and polarimetric results, together with constraints from radio jet observations, suggest a radially extended corona within a truncated disc for Cyg X-1, Swift J1727.8-1613, IGR J17091-3624, and GX 339-4, whereas the disc-corona geometry remains unconstrained for 4U 1957+115, LMC X-3, and 4U 1630-47. We discuss the implications of these findings for understanding accretion geometries and highlight prospects for future X-ray polarimetric studies. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:38:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Majumder",
"Seshadri",
"",
"IITG"
],
[
"Kushwaha",
"Ankur",
"",
"URSC"
],
[
"Singh",
"Swapnil",
"",
"URSC"
],
[
"Jayasurya",
"Kiran M.",
"",
"URSC"
],
[
"Das",
"Santabrata",
"",
"IITG"
],
[
"Nandi",
"Anuj",
"",
"URSC"
]
] |
2506.03775 | Arttu Arjas | Arttu Arjas and Italo Atzeni | Nonlinear Sparse Bayesian Learning Methods with Application to Massive MIMO Channel Estimation with Hardware Impairments | 13 pages, 9 figures | null | null | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | Accurate channel estimation is critical for realizing the performance gains of massive multiple-input multiple-output (MIMO) systems. Traditional approaches to channel estimation typically assume ideal receiver hardware and linear signal models. However, practical receivers suffer from impairments such as nonlinearities in the low-noise amplifiers and quantization errors, which invalidate standard model assumptions and degrade the estimation accuracy. In this work, we propose a nonlinear channel estimation framework that models the distortion function arising from hardware impairments using Gaussian process (GP) regression while leveraging the inherent sparsity of massive MIMO channels. First, we form a GP-based surrogate of the distortion function, employing pseudo-inputs to reduce the computational complexity. Then, we integrate the GP-based surrogate of the distortion function into newly developed enhanced sparse Bayesian learning (SBL) methods, enabling distortion-aware sparse channel estimation. Specifically, we propose two nonlinear SBL methods based on distinct optimization objectives, each offering a different trade-off between estimation accuracy and computational complexity. Numerical results demonstrate significant gains over the Bussgang linear minimum mean squared error estimator and linear SBL, particularly under strong distortion and at high signal-to-noise ratio. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:38:59 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Arjas",
"Arttu",
""
],
[
"Atzeni",
"Italo",
""
]
] |
2506.03776 | Matteo Talluri | Chiara Gambicchia, Enzo Maria Merlino, Berardo Ruffini and Matteo Talluri | Barycentric stability of nonlocal perimeters: the convex case | 12 pages | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | In this work, we establish a sharp form of a nonlocal quantitative isoperimetric inequality involving the barycentric asymmetry for convex sets. This result can be seen as the nonlocal analogue of the one obtained by Fuglede in the 1993. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:39:04 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gambicchia",
"Chiara",
""
],
[
"Merlino",
"Enzo Maria",
""
],
[
"Ruffini",
"Berardo",
""
],
[
"Talluri",
"Matteo",
""
]
] |
2506.03777 | Li Zhang | Li Zhang, Zhongxuan Han, Chaochao chen, Xiaohua Feng, Jiaming Zhang, Yuyuan Li | FedFACT: A Provable Framework for Controllable Group-Fairness Calibration in Federated Learning | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With emerging application of Federated Learning (FL) in decision-making scenarios, it is imperative to regulate model fairness to prevent disparities across sensitive groups (e.g., female, male). Current research predominantly focuses on two concepts of group fairness within FL: Global Fairness (overall model disparity across all clients) and Local Fairness (the disparity within each client). However, the non-decomposable, non-differentiable nature of fairness criteria pose two fundamental, unresolved challenges for fair FL: (i) Harmonizing global and local fairness in multi-class classification; (ii) Enabling a controllable, optimal accuracy-fairness trade-off. To tackle the aforementioned challenges, we propose a novel controllable federated group-fairness calibration framework, named FedFACT. FedFACT identifies the Bayes-optimal classifiers under both global and local fairness constraints in multi-class case, yielding models with minimal performance decline while guaranteeing fairness. To effectively realize an adjustable, optimal accuracy-fairness balance, we derive specific characterizations of the Bayes-optimal fair classifiers for reformulating fair FL as personalized cost-sensitive learning problem for in-processing, and bi-level optimization for post-processing. Theoretically, we provide convergence and generalization guarantees for FedFACT to approach the near-optimal accuracy under given fairness levels. Extensive experiments on multiple datasets across various data heterogeneity demonstrate that FedFACT consistently outperforms baselines in balancing accuracy and global-local fairness. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:39:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zhang",
"Li",
""
],
[
"Han",
"Zhongxuan",
""
],
[
"chen",
"Chaochao",
""
],
[
"Feng",
"Xiaohua",
""
],
[
"Zhang",
"Jiaming",
""
],
[
"Li",
"Yuyuan",
""
]
] |
2506.03778 | Adrita Chakraborty | Adrita Chakraborty | Notes on exact solvability for rotating and pulsating strings in nonrelativistic Lifshitz background | 1+19 pages, 3 figures, submitted to the European Physical journal C | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct one dimensional exactly solvable model by choosing a probe fundamental string rotating and pulsating in the planar Lifshitz spacetime that follows nonrelativistic Lifshitz scaling. We present suitable sets of embedding coordinates for rotating and pulsating strings to embed the string worldsheet on a hyperboloid with anisotropy-dependent eccentricity. The resulting worldsheet Lagrangians straightforwardly reduce to the Lagrangian of a Neumann-Rosochatius integrable model. Although the model assumes exact solutions for both the chosen ansatz its classical Liouville integrability is found to be conditional due to the presence of finite anisotropy in the target space geometry. We further use the exact solutions of the model to yield energy-momentum dispersion relations. We interpret those from the perspective of highly degenerate frustrated $J_1-J_2$ spin chain for rotating string and frustration-free Motzkin spin chain for pulsating string. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:40:47 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chakraborty",
"Adrita",
""
]
] |
2506.03779 | Hachem Kadri | Hachem Kadri, Joachim Tomasi, Yuka Hashimoto, Sandrine Anthoine | Towards Quantum Operator-Valued Kernels | null | null | null | null | quant-ph cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Quantum kernels are reproducing kernel functions built using quantum-mechanical principles and are studied with the aim of outperforming their classical counterparts. The enthusiasm for quantum kernel machines has been tempered by recent studies that have suggested that quantum kernels could not offer speed-ups when learning on classical data. However, most of the research in this area has been devoted to scalar-valued kernels in standard classification or regression settings for which classical kernel methods are efficient and effective, leaving very little room for improvement with quantum kernels. This position paper argues that quantum kernel research should focus on more expressive kernel classes. We build upon recent advances in operator-valued kernels, and propose guidelines for investigating quantum kernels. This should help to design a new generation of quantum kernel machines and fully explore their potentials. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:40:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Kadri",
"Hachem",
""
],
[
"Tomasi",
"Joachim",
""
],
[
"Hashimoto",
"Yuka",
""
],
[
"Anthoine",
"Sandrine",
""
]
] |
2506.03780 | Hasan Fallahgoul | Hasan Fallahgoul | High-Dimensional Learning in Finance | null | null | null | null | q-fin.ST cs.LG econ.EM stat.ML | http://creativecommons.org/licenses/by/4.0/ | Recent advances in machine learning have shown promising results for financial prediction using large, over-parameterized models. This paper provides theoretical foundations and empirical validation for understanding when and how these methods achieve predictive success. I examine three key aspects of high-dimensional learning in finance. First, I prove that within-sample standardization in Random Fourier Features implementations fundamentally alters the underlying Gaussian kernel approximation, replacing shift-invariant kernels with training-set dependent alternatives. Second, I derive sample complexity bounds showing when reliable learning becomes information-theoretically impossible under weak signal-to-noise ratios typical in finance. Third, VC-dimension analysis reveals that ridgeless regression's effective complexity is bounded by sample size rather than nominal feature dimension. Comprehensive numerical validation confirms these theoretical predictions, revealing systematic breakdown of claimed theoretical properties across realistic parameter ranges. These results show that when sample size is small and features are high-dimensional, observed predictive success is necessarily driven by low-complexity artifacts, not genuine high-dimensional learning. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:41:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fallahgoul",
"Hasan",
""
]
] |
2506.03781 | Seungcheol Park | Seungcheol Park, Jeongin Bae, Beomseok Kwon, Minjun Kim, Byeongwook Kim, Se Jung Kwon, U Kang, Dongsoo Lee | Unifying Uniform and Binary-coding Quantization for Accurate Compression of Large Language Models | ACL 2025 Main Track | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | How can we quantize large language models while preserving accuracy? Quantization is essential for deploying large language models (LLMs) efficiently. Binary-coding quantization (BCQ) and uniform quantization (UQ) are promising quantization schemes that have strong expressiveness and optimizability, respectively. However, neither scheme leverages both advantages. In this paper, we propose UniQuanF (Unified Quantization with Flexible Mapping), an accurate quantization method for LLMs. UniQuanF harnesses both strong expressiveness and optimizability by unifying the flexible mapping technique in UQ and non-uniform quantization levels of BCQ. We propose unified initialization, and local and periodic mapping techniques to optimize the parameters in UniQuanF precisely. After optimization, our unification theorem removes computational and memory overhead, allowing us to utilize the superior accuracy of UniQuanF without extra deployment costs induced by the unification. Experimental results demonstrate that UniQuanF outperforms existing UQ and BCQ methods, achieving up to 4.60% higher accuracy on GSM8K benchmark. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:42:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Park",
"Seungcheol",
""
],
[
"Bae",
"Jeongin",
""
],
[
"Kwon",
"Beomseok",
""
],
[
"Kim",
"Minjun",
""
],
[
"Kim",
"Byeongwook",
""
],
[
"Kwon",
"Se Jung",
""
],
[
"Kang",
"U",
""
],
[
"Lee",
"Dongsoo",
""
]
] |
2506.03782 | Raghunath Sahoo | Suraj Prasad, Sushanta Tripathy, Bhagyarathi Sahoo, and Raghunath Sahoo | Event Topology Classifiers at the Large Hadron Collider | 68 pages and 82 captioned figures. Submitted for publication. Comments and Suggestions are most welcome | null | null | null | hep-ph hep-ex hep-th nucl-ex nucl-th | http://creativecommons.org/licenses/by-sa/4.0/ | Event classifiers are the most fundamental observables to probe the system size and event topology of high-energy collisions. Significant progress has been made to establish suitable event classifiers to probe different physics processes occurring in elementary $e^{+}e^{-}$ to heavy-ion collisions in a range of center of mass energies. One of the major motivations to revisit event classifiers at the Large Hadron Collider (LHC) originates from the recent measurements of high multiplicity proton-proton collisions, which have revealed that these small collision systems exhibit features similar to the formation of quark-gluon plasma (QGP), traditionally believed to be only achievable in heavy nucleus-nucleus collisions at ultra-relativistic energies. To pinpoint the origin of these QGP-like phenomena with substantially reduced auto-correlation and selection biases, and to bring all collision systems on equal footing, along with charged-particle multiplicity, lately several event topology classifiers such as transverse sphericity, transverse spherocity, relative transverse activity classifier, and charged-particle flattenicity have been used extensively in experiments as well as in the phenomenological front. In this review article, we extensively discuss the purpose, coverage, and usage guidelines of these event shape classifiers, which can be crucial given the upcoming measurements in a precision era during Runs 3 and 4 of the LHC. In most cases, the event shape observables are found to be better probes in understanding the heavy-ion-like behavior seen at the LHC, while making a multi-differential study of multihadron production dynamics in hadronic and nuclear collisions. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:43:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Prasad",
"Suraj",
""
],
[
"Tripathy",
"Sushanta",
""
],
[
"Sahoo",
"Bhagyarathi",
""
],
[
"Sahoo",
"Raghunath",
""
]
] |
2506.03783 | Itamar Oliveira | Jonathan Bennett, Neal Bez, Susana Gutierrez, Shohei Nakamura, Itamar Oliveira | Mizohata-Takeuchi inequalities for orthonormal systems | null | null | null | null | math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We establish some weighted $L^2$ inequalities for Fourier extension operators in the setting of orthonormal systems. In the process we develop a direct approach to such inequalities based on generalised Wigner distributions, complementing the Schatten duality approach that is prevalent in the wider context of estimates for such orthonormal systems. Our results are set within a broader family of tentatively suggested ($L^p$) inequalities of Mizohata-Takeuchi type. For $p$ a power of two at least, such weighted inequalities may be recast as questions of co-positivity of tensor forms. For $p\leq 1$ we provide some evidence that they may hold in reverse provided the orthonormal sequence is complete. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:44:17 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bennett",
"Jonathan",
""
],
[
"Bez",
"Neal",
""
],
[
"Gutierrez",
"Susana",
""
],
[
"Nakamura",
"Shohei",
""
],
[
"Oliveira",
"Itamar",
""
]
] |
2506.03784 | Emanuele Marconato | Beatrix M. G. Nielsen, Emanuele Marconato, Andrea Dittadi, Luigi Gresele | When Does Closeness in Distribution Imply Representational Similarity? An Identifiability Perspective | null | null | null | null | cs.LG cs.AI stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | When and why representations learned by different deep neural networks are similar is an active research topic. We choose to address these questions from the perspective of identifiability theory, which suggests that a measure of representational similarity should be invariant to transformations that leave the model distribution unchanged. Focusing on a model family which includes several popular pre-training approaches, e.g., autoregressive language models, we explore when models which generate distributions that are close have similar representations. We prove that a small Kullback-Leibler divergence between the model distributions does not guarantee that the corresponding representations are similar. This has the important corollary that models arbitrarily close to maximizing the likelihood can still learn dissimilar representations, a phenomenon mirrored in our empirical observations on models trained on CIFAR-10. We then define a distributional distance for which closeness implies representational similarity, and in synthetic experiments, we find that wider networks learn distributions which are closer with respect to our distance and have more similar representations. Our results establish a link between closeness in distribution and representational similarity. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:44:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nielsen",
"Beatrix M. G.",
""
],
[
"Marconato",
"Emanuele",
""
],
[
"Dittadi",
"Andrea",
""
],
[
"Gresele",
"Luigi",
""
]
] |
2506.03786 | Lulu Nie | Lulu Nie, Baoqin Wen, Jingbin Li, Shufeng Li, Yali Li, Zhaokun Zhang, Zhiyuan Wang, Zhihao Fan | Discrete Element Parameter Calibration of Livestock Salt Based on Particle Scaling | null | null | null | null | eess.SY cs.SY physics.med-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In order to obtain accurate contact parameters for the discrete element simulation of salt particles used in animal husbandry, the principle of particle contact scaling and dimensional analysis were used for particle scaling. Firstly, the Plackett Burman experiment was used to screen the parameters that significantly affect the angle of repose: salt salt rolling friction coefficient, salt salt recovery coefficient, and salt steel rolling friction coefficient. Considering the influence of other parameters, a combination of bench and simulation experiments was used to calibrate the contact parameters between salt particles and steel plates used in animal husbandry in EDEM. Finally, through the stacking test, steepest climbing test, and orthogonal rotation combination test, the salt salt rolling friction coefficient was obtained to be 0.23, the salt salt recovery coefficient was 0.544, and the salt steel rolling friction coefficient was 0.368, which were verified through bench tests. The experimental results show that the relative error between the actual value of the stacking angle and the simulation results is 0.6%. The results indicate that the calibrated contact parameters can be used for discrete element simulation of salt particles for animal husbandry, providing reference for the design of quantitative feeding screws and silos. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:47:07 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nie",
"Lulu",
""
],
[
"Wen",
"Baoqin",
""
],
[
"Li",
"Jingbin",
""
],
[
"Li",
"Shufeng",
""
],
[
"Li",
"Yali",
""
],
[
"Zhang",
"Zhaokun",
""
],
[
"Wang",
"Zhiyuan",
""
],
[
"Fan",
"Zhihao",
""
]
] |
2506.03787 | Claudio Ferrari Dr | Elena Ferrari, Claudio Ferrari, Jos\`e Baruchel, Thu Ni Tran Caliste, Stefano Basso, Akanshu Chauhan, Bianca Salmaso, Davide Sisana, Daniele Spiga | Effect of chromatic dispersion on image size and lattice plane curvature measurements with Rocking Curve Imaging | 9 pages, 4 figures | null | null | null | physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | ADP crystals of large dimensions (80x80x20 mm3), to be used as a X-ray beam expanders in the BEaTriX facility at INAF-OABrera, have been characterised at BM05 beamline at ESRF synchrotron with the main purpose to determine lattice plane curvature with a unprecedent accuracy, as the BEaTriX setup requires a radius of curvature larger than 22 km. In this beamline, the monochromator is made by 2 Si(111) parallel crystals in the non-dispersive configuration. Due to the difference in the Bragg angles between the Si(111) monochromator and the ADP(008) diffractions, only a limited part of the sample area, hit by the X-ray beam (11x11 mm2 cross section), produced a diffracted beam for a given value of angle of incidence. In the rocking curve imaging techniques, a full image of the sample for a given peak position is obtained by combining images taken at different angles of incidence compensated by the dispersion correction, that is the Bragg condition difference in different points of the sample. It is found that the resulting image size parallel to the scattering plane is affected by this dispersion correction. A 4.5 % elongation along the scattering plane was evaluated in the present Bragg case. This contribution is opposite in the Laue case. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:47:28 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ferrari",
"Elena",
""
],
[
"Ferrari",
"Claudio",
""
],
[
"Baruchel",
"Josè",
""
],
[
"Caliste",
"Thu Ni Tran",
""
],
[
"Basso",
"Stefano",
""
],
[
"Chauhan",
"Akanshu",
""
],
[
"Salmaso",
"Bianca",
""
],
[
"Sisana",
"Davide",
""
],
[
"Spiga",
"Daniele",
""
]
] |
2506.03788 | Chiara Boldrini | Kamer Cekini and Elisabetta Biondi and Chiara Boldrini and Andrea Passarella and Marco Conti | The Impact of COVID-19 on Twitter Ego Networks: Structure, Sentiment, and Topics | Funding: SoBigData.it (IR0000013), SoBigData PPP (101079043), FAIR (PE00000013), SERICS (PE00000014), ICSC (CN00000013) | null | null | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lockdown measures, implemented by governments during the initial phases of the COVID-19 pandemic to reduce physical contact and limit viral spread, imposed significant restrictions on in-person social interactions. Consequently, individuals turned to online social platforms to maintain connections. Ego networks, which model the organization of personal relationships according to human cognitive constraints on managing meaningful interactions, provide a framework for analyzing such dynamics. The disruption of physical contact and the predominant shift of social life online potentially altered the allocation of cognitive resources dedicated to managing these digital relationships. This research aims to investigate the impact of lockdown measures on the characteristics of online ego networks, presumably resulting from this reallocation of cognitive resources. To this end, a large dataset of Twitter users was examined, covering a seven-year period of activity. Analyzing a seven-year Twitter dataset -- including five years pre-pandemic and two years post -- we observe clear, though temporary, changes. During lockdown, ego networks expanded, social circles became more structured, and relationships intensified. Simultaneously, negative interactions increased, and users engaged with a broader range of topics, indicating greater thematic diversity. Once restrictions were lifted, these structural, emotional, and thematic shifts largely reverted to pre-pandemic norms -- suggesting a temporary adaptation to an extraordinary social context. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:48:29 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Cekini",
"Kamer",
""
],
[
"Biondi",
"Elisabetta",
""
],
[
"Boldrini",
"Chiara",
""
],
[
"Passarella",
"Andrea",
""
],
[
"Conti",
"Marco",
""
]
] |
2506.03789 | Ruediger Klingeler | Luca Bischof and Jan Arneth and Raju Kalaivanan and Raman Sankar and Kwang-Yong Choi and R\"udiger Klingeler | Spin waves in Na$_2$Co$_2$TeO$_6$ studied by high-frequency/high-field ESR: Successes and failures of the triple-$\mathbf{q}$ model | null | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The Kitaev candidate material Na$_2$Co$_2$TeO$_6$ is proposed to be proximate to a quantum spin liquid state but a suitable spin model and the nature of its ground states are still under debate. Our high-frequency/high-field electron spin resonance spectroscopy studies of Na$_2$Co$_2$TeO$_6$ single-crystals under in-plane and out-of-plane magnetic fields elucidate the ground state by investigating its low-energy spin wave excitations. Several excitation modes are observed in the low-field phase and in the phases induced by $B\parallel a^*$. In addition, the spectra exhibit a frequency-independent feature at the phase boundary connected to the putative quantum phase transition. For magnetic fields applied along the $c$ axis, the observation of three distinct spin wave modes in the antiferromagnetic (AFM) ground state reveals a previously unresolved splitting of the zero-field excitation gap into $\Delta = 211\,$GHz and $\Delta_2 = 237\,$GHz. The softening of one of these modes evidences a field-induced phase transition at $B_{\rm c1} = 4.7\,$T, which is corroborated by a clear anomaly in the isothermal magnetization. Spin wave calculations based on the extended Heisenberg-Kitaev model exclude a zigzag ground state of the AFM phase. A triple-q spin configuration correctly predicts two spin wave modes, but fails to reproduce the softening mode. Our analysis shows that the triple-q ground state model of Na$_2$Co$_2$TeO$_6$ is incomplete and suggests the relevance of interlayer interactions. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:49:20 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bischof",
"Luca",
""
],
[
"Arneth",
"Jan",
""
],
[
"Kalaivanan",
"Raju",
""
],
[
"Sankar",
"Raman",
""
],
[
"Choi",
"Kwang-Yong",
""
],
[
"Klingeler",
"Rüdiger",
""
]
] |
2506.03790 | Peng Wang | Peng Wang, Yifu Lu, Yaodong Yu, Druv Pai, Qing Qu, Yi Ma | Attention-Only Transformers via Unrolled Subspace Denoising | 28 pages, 7 figures, 5 tables | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite the popularity of transformers in practice, their architectures are empirically designed and neither mathematically justified nor interpretable. Moreover, as indicated by many empirical studies, some components of transformer architectures may be redundant. To derive a fully interpretable transformer architecture with only necessary components, we contend that the goal of representation learning is to compress a set of noisy initial token representations towards a mixture of low-dimensional subspaces. To compress these noisy token representations, an associated denoising operation naturally takes the form of a multi-head (subspace) self-attention. By unrolling such iterative denoising operations into a deep network, we arrive at a highly compact architecture that consists of \textit{only} self-attention operators with skip connections at each layer. Moreover, we show that each layer performs highly efficient denoising: it improves the signal-to-noise ratio of token representations \textit{at a linear rate} with respect to the number of layers. Despite its simplicity, extensive experiments on vision and language tasks demonstrate that such a transformer achieves performance close to that of standard transformer architectures such as GPT-2 and CRATE. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:53:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wang",
"Peng",
""
],
[
"Lu",
"Yifu",
""
],
[
"Yu",
"Yaodong",
""
],
[
"Pai",
"Druv",
""
],
[
"Qu",
"Qing",
""
],
[
"Ma",
"Yi",
""
]
] |
2506.03791 | Michal Nowak | M. P. Nowak, M. Zegrodnik, D. Grzelec, B. Szafran, R. Citro, P. W\'ojcik | Scanning gate microscopy probing of anisotropic electron flow in a two dimensional electron gas at the (110) $\mathrm{LaAlO}_3/\mathrm{SrTiO}_3$ interface | null | null | null | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | We theoretically investigate the anisotropic dispersion features of a two dimensional electron gas at the (110) oriented $\mathrm{LaAlO}_3/\mathrm{SrTiO}_3$ interfaces, as revealed by scanning gate microscopy of electronic flow from a quantum point contact. The dispersion relation of the (110) $\mathrm{LaAlO}_3/\mathrm{SrTiO}_3$ interface is characterized by a highly non-circular Fermi surface. Here, we develop an efficient tight-binding model for the electron gas at the interface. We show that the anisotropy of the Fermi surface causes both the direction of the electron flux from the quantum point contact and the periodicity of the self-interference conductance fringes to depend strongly on the orientation of the constriction relative to the crystal lattice. We show that the radially non-uniform distribution of the Fermi velocity on the Fermi surface results in skewing of electron trajectories when the quantum point contact gates are not aligned with the in-plane primitive vectors. We show that this effect results in the separation of electrons belonging to different orbitals for wide (110) $\mathrm{LaAlO}_3/\mathrm{SrTiO}_3$ quantum wells. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:54:01 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nowak",
"M. P.",
""
],
[
"Zegrodnik",
"M.",
""
],
[
"Grzelec",
"D.",
""
],
[
"Szafran",
"B.",
""
],
[
"Citro",
"R.",
""
],
[
"Wójcik",
"P.",
""
]
] |
2506.03792 | Qianwei Qu | Qianwei Qu, Christian M. Schlep\"utz, Marco Stampanoni | Analytical Reconstruction of Periodically Deformed Objects in Time-resolved CT | null | null | null | null | physics.med-ph cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time-resolved CT is an advanced measurement technique that has been widely used to observe dynamic objects, including periodically varying structures such as hearts, lungs, or hearing structures. To reconstruct these objects from CT projections, a common approach is to divide the projections into several collections based on their motion phases and perform reconstruction within each collection, assuming they originate from a static object. This describes the gating-based method, which is the standard approach for time-periodic reconstruction. However, the gating-based reconstruction algorithm only utilizes a limited subset of projections within each collection and ignores the correlation between different collections, leading to inefficient use of the radiation dose. To address this issue, we propose two analytical reconstruction pipelines in this paper, and validate them with experimental data captured using tomographic synchrotron microscopy. We demonstrate that our approaches significantly reduce random noise in the reconstructed images without blurring the sharp features of the observed objects. Equivalently, our methods can achieve the same reconstruction quality as gating-based methods but with a lower radiation dose. Our code is available at github.com/PeriodRecon. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:54:30 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Qu",
"Qianwei",
""
],
[
"Schlepütz",
"Christian M.",
""
],
[
"Stampanoni",
"Marco",
""
]
] |
2506.03793 | Sparsh Jain | Sidharth Pulipaka, Sparsh Jain, Ashwin Sankar, Raj Dabre | Mark My Words: A Robust Multilingual Model for Punctuation in Text and Speech Transcripts | Work in Progress | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Punctuation plays a vital role in structuring meaning, yet current models often struggle to restore it accurately in transcripts of spontaneous speech, especially in the presence of disfluencies such as false starts and backtracking. These limitations hinder the performance of downstream tasks like translation, text to speech, summarization, etc. where sentence boundaries are critical for preserving quality. In this work, we introduce Cadence, a generalist punctuation restoration model adapted from a pretrained large language model. Cadence is designed to handle both clean written text and highly spontaneous spoken transcripts. It surpasses the previous state of the art in performance while expanding support from 14 to all 22 Indian languages and English. We conduct a comprehensive analysis of model behavior across punctuation types and language families, identifying persistent challenges under domain shift and with rare punctuation marks. Our findings demonstrate the efficacy of utilizing pretrained language models for multilingual punctuation restoration and highlight Cadence practical value for low resource NLP pipelines at scale. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:54:38 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pulipaka",
"Sidharth",
""
],
[
"Jain",
"Sparsh",
""
],
[
"Sankar",
"Ashwin",
""
],
[
"Dabre",
"Raj",
""
]
] |
2506.03794 | Onur Baysal | Onur Baysal, Maria Aquilina | An Efficient Numerical Method for an Approximate Solution of the Beam Equation | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we propose a horizontal type method of lines numerical scheme for the unsteady Euler-Bernoulli beam equation. The problem is initially reformulated as a first order system of initial value problems and a suitable one-step difference scheme is used for the highest order temporal derivative which leads to a system of steady beam equations. Then resulted family of steady problems is solved iteratively by the finite element method with Hermite cubic basis functions. This iterative procedure leads to approximations for both the solution of the unsteady problem and its derivatives. All these approximations are compared with the exact ones to illustrate the performance of the proposed method. Moreover, the optimization of the mesh parameters is discussed for both steady and unsteady problems by logarithmic scale plot. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:55:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Baysal",
"Onur",
""
],
[
"Aquilina",
"Maria",
""
]
] |
2506.03795 | Ludovic Dumoulin | Ludovic Dumoulin, Carles Blanch-Mercader and Karsten Kruse | Defect states in compressible active polar fluids with turnover | null | null | null | null | physics.bio-ph cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | Biological active matter like the cytoskeleton or tissues are characterized by their ability to transform chemical energy into mechanical stress. In addition, it often exhibits orientational order, which is essential for many cellular and morphogenetic processes. Experimental evidence suggests that defects in the orientational order field play an important role in organizing active stress. However, defects tend to annihilate unless the material is in a chaotic state or hydrodynamic interactions are suppressed. Using a hydrodynamic description of compressible active polar fluids, we show that turnover readily leads to a stabilization of defects. Depending on the turnover rate, topological defects arrange in a multitude of different phases, including lattices, active foams, and vortex glasses. Our work suggests that turnover plays a crucial role for organizing biological active matter. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:56:08 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Dumoulin",
"Ludovic",
""
],
[
"Blanch-Mercader",
"Carles",
""
],
[
"Kruse",
"Karsten",
""
]
] |
2506.03796 | Penny Madysa | Penelope Madysa and Sabrina Appel and Verena Kain and Michael Schenk | Geoff: The Generic Optimization Framework & Frontend for Particle Accelerator Controls | 18 pages, 5 figures. Submitted to SoftwareX | null | null | null | physics.acc-ph cs.LG | http://creativecommons.org/licenses/by/4.0/ | Geoff is a collection of Python packages that form a framework for automation of particle accelerator controls. With particle accelerator laboratories around the world researching machine learning techniques to improve accelerator performance and uptime, a multitude of approaches and algorithms have emerged. The purpose of Geoff is to harmonize these approaches and to minimize friction when comparing or migrating between them. It provides standardized interfaces for optimization problems, utility functions to speed up development, and a reference GUI application that ties everything together. Geoff is an open-source library developed at CERN and maintained and updated in collaboration between CERN and GSI as part of the EURO-LABS project. This paper gives an overview over Geoff's design, features, and current usage. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 09:58:48 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Madysa",
"Penelope",
""
],
[
"Appel",
"Sabrina",
""
],
[
"Kain",
"Verena",
""
],
[
"Schenk",
"Michael",
""
]
] |
2506.03797 | Khalil Teber | Khalil Teber, M\'elanie Weynants, Fabian Gans, Miguel D. Mahecha | Geo-Disasters: geocoding climate-related events in the international disaster database EM-DAT | 16 pages, 7 figures, 1 table. Correspondence to: [email protected]. Data available at DOI: https://doi.org/10.5281/zenodo.15487667 | null | null | null | physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | Climate hazards can escalate into humanitarian disasters. Understanding their trajectories -- considering hazard intensity, human exposure, and societal vulnerability -- is essential for effective anticipatory action. The International Disaster Database (EM-DAT) is the only freely available global resource of humanitarian disaster records. However, it lacks exact geospatial information, limiting its use for climate hazard impact research. Here, we provide geocoding of 9,217 climate-related disasters reported by EM-DAT from 1990 to 2023, along with an open, reproducible framework for updating. Our method remains accurate even when only region names are available and includes quality flags to assess reliability. The augmented EM-DAT enables integration with other geocoded data, supporting more accurate assessment of climate disaster impacts and adaptation deficits. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:00:05 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Teber",
"Khalil",
""
],
[
"Weynants",
"Mélanie",
""
],
[
"Gans",
"Fabian",
""
],
[
"Mahecha",
"Miguel D.",
""
]
] |
2506.03798 | Fan Shi | Fan Shi, Haiyang Yu, Bin Li, Xiangyang Xue | CoLa: Chinese Character Decomposition with Compositional Latent Components | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humans can decompose Chinese characters into compositional components and recombine them to recognize unseen characters. This reflects two cognitive principles: Compositionality, the idea that complex concepts are built on simpler parts; and Learning-to-learn, the ability to learn strategies for decomposing and recombining components to form new concepts. These principles provide inductive biases that support efficient generalization. They are critical to Chinese character recognition (CCR) in solving the zero-shot problem, which results from the common long-tail distribution of Chinese character datasets. Existing methods have made substantial progress in modeling compositionality via predefined radical or stroke decomposition. However, they often ignore the learning-to-learn capability, limiting their ability to generalize beyond human-defined schemes. Inspired by these principles, we propose a deep latent variable model that learns Compositional Latent components of Chinese characters (CoLa) without relying on human-defined decomposition schemes. Recognition and matching can be performed by comparing compositional latent components in the latent space, enabling zero-shot character recognition. The experiments illustrate that CoLa outperforms previous methods in both character the radical zero-shot CCR. Visualization indicates that the learned components can reflect the structure of characters in an interpretable way. Moreover, despite being trained on historical documents, CoLa can analyze components of oracle bone characters, highlighting its cross-dataset generalization ability. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:06:15 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Shi",
"Fan",
""
],
[
"Yu",
"Haiyang",
""
],
[
"Li",
"Bin",
""
],
[
"Xue",
"Xiangyang",
""
]
] |
2506.03799 | Fei Zhang | Fei Zhang, Pei Zhang, Baosong Yang, Fei Huang, Yanfeng Wang, Ya Zhang | ConText: Driving In-context Learning for Text Removal and Segmentation | 19 pages, 9 figures, Accepted at ICML 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper presents the first study on adapting the visual in-context learning (V-ICL) paradigm to optical character recognition tasks, specifically focusing on text removal and segmentation. Most existing V-ICL generalists employ a reasoning-as-reconstruction approach: they turn to using a straightforward image-label compositor as the prompt and query input, and then masking the query label to generate the desired output. This direct prompt confines the model to a challenging single-step reasoning process. To address this, we propose a task-chaining compositor in the form of image-removal-segmentation, providing an enhanced prompt that elicits reasoning with enriched intermediates. Additionally, we introduce context-aware aggregation, integrating the chained prompt pattern into the latent query representation, thereby strengthening the model's in-context reasoning. We also consider the issue of visual heterogeneity, which complicates the selection of homogeneous demonstrations in text recognition. Accordingly, this is effectively addressed through a simple self-prompting strategy, preventing the model's in-context learnability from devolving into specialist-like, context-free inference. Collectively, these insights culminate in our ConText model, which achieves new state-of-the-art across both in- and out-of-domain benchmarks. The code is available at https://github.com/Ferenas/ConText. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:06:32 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zhang",
"Fei",
""
],
[
"Zhang",
"Pei",
""
],
[
"Yang",
"Baosong",
""
],
[
"Huang",
"Fei",
""
],
[
"Wang",
"Yanfeng",
""
],
[
"Zhang",
"Ya",
""
]
] |
2506.03800 | Wenjun Lin | Hongwang Xiao, Wenjun Lin, Xi Chen, Hui Wang, Kai Chen, Jiashan Li, Yuancheng Sun, Sicheng Dai, Boya Wu, Qiwei Ye | STELLA: Towards Protein Function Prediction with Multimodal LLMs Integrating Sequence-Structure Representations | null | null | null | null | q-bio.BM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Protein biology focuses on the intricate relationships among sequences, structures, and functions. Deciphering protein functions is crucial for understanding biological processes, advancing drug discovery, and enabling synthetic biology applications. Since protein sequences determine tertiary structures, which in turn govern functions, integrating sequence and structure information is essential for accurate prediction of protein functions. Traditional protein language models (pLMs) have advanced protein-related tasks by learning representations from large-scale sequence and structure data. However, pLMs are limited in integrating broader contextual knowledge, particularly regarding functional modalities that are fundamental to protein biology. In contrast, large language models (LLMs) have exhibited outstanding performance in contextual understanding, reasoning, and generation across diverse domains. Leveraging these capabilities, STELLA is proposed as a multimodal LLM integrating protein sequence-structure representations with general knowledge to address protein function prediction. Through multimodal instruction tuning (MMIT) using the proposed OPI-Struc dataset, STELLA achieves state-of-the-art performance in two function-related tasks-functional description prediction (FP) and enzyme-catalyzed reaction prediction (EP). This study highlights the potential of multimodal LLMs as an alternative paradigm to pLMs to advance protein biology research. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:09:15 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xiao",
"Hongwang",
""
],
[
"Lin",
"Wenjun",
""
],
[
"Chen",
"Xi",
""
],
[
"Wang",
"Hui",
""
],
[
"Chen",
"Kai",
""
],
[
"Li",
"Jiashan",
""
],
[
"Sun",
"Yuancheng",
""
],
[
"Dai",
"Sicheng",
""
],
[
"Wu",
"Boya",
""
],
[
"Ye",
"Qiwei",
""
]
] |
2506.03801 | Nijat Mehdiyev Dr. | Peter Pfeiffer, Alexander Rombach, Maxim Majlatow and Nijat Mehdiyev | From Theory to Practice: Real-World Use Cases on Trustworthy LLM-Driven Process Modeling, Prediction and Automation | Accepted to the Next Gen Data and Process Management: Large Language Models and Beyond workshop at SIGMOD 2025 | null | null | null | cs.SE cs.LG cs.MA | http://creativecommons.org/licenses/by/4.0/ | Traditional Business Process Management (BPM) struggles with rigidity, opacity, and scalability in dynamic environments while emerging Large Language Models (LLMs) present transformative opportunities alongside risks. This paper explores four real-world use cases that demonstrate how LLMs, augmented with trustworthy process intelligence, redefine process modeling, prediction, and automation. Grounded in early-stage research projects with industrial partners, the work spans manufacturing, modeling, life-science, and design processes, addressing domain-specific challenges through human-AI collaboration. In manufacturing, an LLM-driven framework integrates uncertainty-aware explainable Machine Learning (ML) with interactive dialogues, transforming opaque predictions into auditable workflows. For process modeling, conversational interfaces democratize BPMN design. Pharmacovigilance agents automate drug safety monitoring via knowledge-graph-augmented LLMs. Finally, sustainable textile design employs multi-agent systems to navigate regulatory and environmental trade-offs. We intend to examine tensions between transparency and efficiency, generalization and specialization, and human agency versus automation. By mapping these trade-offs, we advocate for context-sensitive integration prioritizing domain needs, stakeholder values, and iterative human-in-the-loop workflows over universal solutions. This work provides actionable insights for researchers and practitioners aiming to operationalize LLMs in critical BPM environments. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:12:09 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Pfeiffer",
"Peter",
""
],
[
"Rombach",
"Alexander",
""
],
[
"Majlatow",
"Maxim",
""
],
[
"Mehdiyev",
"Nijat",
""
]
] |
2506.03802 | Andreas Athanasopoulos | Andreas Athanasopoulos, Christos Dimitrakakis | Learning Equilibria in Matching Games with Bandit Feedback | 21 pages, 2 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We investigate the problem of learning an equilibrium in a generalized two-sided matching market, where agents can adaptively choose their actions based on their assigned matches. Specifically, we consider a setting in which matched agents engage in a zero-sum game with initially unknown payoff matrices, and we explore whether a centralized procedure can learn an equilibrium from bandit feedback. We adopt the solution concept of matching equilibrium, where a pair consisting of a matching $\mathfrak{m}$ and a set of agent strategies $X$ forms an equilibrium if no agent has the incentive to deviate from $(\mathfrak{m}, X)$. To measure the deviation of a given pair $(\mathfrak{m}, X)$ from the equilibrium pair $(\mathfrak{m}^\star, X^\star)$, we introduce matching instability that can serve as a regret measure for the corresponding learning problem. We then propose a UCB algorithm in which agents form preferences and select actions based on optimistic estimates of the game payoffs, and prove that it achieves sublinear, instance-independent regret over a time horizon $T$. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:15:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Athanasopoulos",
"Andreas",
""
],
[
"Dimitrakakis",
"Christos",
""
]
] |
2506.03803 | Xianfeng Chen | Xianfeng Chen, Nirmala Raj, Ruvi Lecamwasam, Mingxi Chen, Christina Yuan Ling Tan, Syed M. Assad, and Ping Koy Lam | Levitated macroscopic rotors with 10 hours of free spin at room temperature | 11 pages, 6 figures | null | null | null | physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | Low-dissipation rotors with large angular momentum are essential for precision sensing and probing macroscopic quantum phenomena. To date, low dissipation can only be achieved for micro-scale rotors. Here, we report a diamagnetically levitated millimeter-scale rotor exhibiting a measured dissipation rate as low as $3.85\,\mu\mathrm{Hz}$ at room temperature, corresponding to a free spinning duration exceeding 10 hours. The rotor is levitated stably over an axisymmetric permanent magnet trap, and can be driven up to 930 RPM using contactless electrostatic actuation in high vacuum. Leveraging its low damping rate and large angular momentum, we realize a precision gyroscope with a measured sensitivity of $6.5 \times 10^{-3}\ \mathrm{^\circ/s}$ and an estimated thermal-limited stability of $5.7 \times 10^{-7}\ \mathrm{^\circ/\sqrt{h}}$. These results establish diamagnetic levitation as a promising room-temperature platform for high-performance gyroscopes. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:18:57 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Xianfeng",
""
],
[
"Raj",
"Nirmala",
""
],
[
"Lecamwasam",
"Ruvi",
""
],
[
"Chen",
"Mingxi",
""
],
[
"Tan",
"Christina Yuan Ling",
""
],
[
"Assad",
"Syed M.",
""
],
[
"Lam",
"Ping Koy",
""
]
] |
2506.03804 | George Webber | George Webber, Alexander Hammers, Andrew P. King, Andrew J. Reader | Personalized MR-Informed Diffusion Models for 3D PET Image Reconstruction | 10 pages, 10 figures | null | null | null | physics.med-ph cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent work has shown improved lesion detectability and flexibility to reconstruction hyperparameters (e.g. scanner geometry or dose level) when PET images are reconstructed by leveraging pre-trained diffusion models. Such methods train a diffusion model (without sinogram data) on high-quality, but still noisy, PET images. In this work, we propose a simple method for generating subject-specific PET images from a dataset of multi-subject PET-MR scans, synthesizing "pseudo-PET" images by transforming between different patients' anatomy using image registration. The images we synthesize retain information from the subject's MR scan, leading to higher resolution and the retention of anatomical features compared to the original set of PET images. With simulated and real [$^{18}$F]FDG datasets, we show that pre-training a personalized diffusion model with subject-specific "pseudo-PET" images improves reconstruction accuracy with low-count data. In particular, the method shows promise in combining information from a guidance MR scan without overly imposing anatomical features, demonstrating an improved trade-off between reconstructing PET-unique image features versus features present in both PET and MR. We believe this approach for generating and utilizing synthetic data has further applications to medical imaging tasks, particularly because patient-specific PET images can be generated without resorting to generative deep learning or large training datasets. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:24:14 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Webber",
"George",
""
],
[
"Hammers",
"Alexander",
""
],
[
"King",
"Andrew P.",
""
],
[
"Reader",
"Andrew J.",
""
]
] |
2506.03805 | Simeon Ball | Simeon Ball and Tabriz Popatia | Additive codes from linear codes | null | null | null | null | cs.IT math.CO math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce two constructions of additive codes over finite fields. Both constructions start with a linear code over a field with $q$ elements and give additive codes over the field with $q^h$ elements whose minimum distance is demonstrably good. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:26:24 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Ball",
"Simeon",
""
],
[
"Popatia",
"Tabriz",
""
]
] |
2506.03806 | Mohamad Nasser | Mohamad N. Nasser, Vaibhav Keshari, and Madeti Prabhakar | Matrix representations of the twisted virtual braid group and its extensions | null | null | null | null | math.RT | http://creativecommons.org/licenses/by/4.0/ | This paper classifies complex local representations of the twisted virtual braid group, $\mathrm{TVB}_2$, into $\mathrm{GL}_3(\mathbb{C})$. It shows that such representations fall into eight types, all of which are unfaithful and reducible to degree $2 \times 2$. Further reducibility to degree 1 is analyzed for specific types. The paper also examines complex homogeneous local representations of $\mathrm{TVB}_n$ into $\mathrm{GL}_{n+1}(\mathbb{C})$ for $n \geq 3$, identifying seven unfaithful types. Additionally, complex local representations of the singular twisted virtual braid group, $\mathrm{STVB}_2$, into $M_3(\mathbb{C})$ are classified into thirteen unfaithful types. Finally, the paper demonstrates that not all complex local extensions of $\mathrm{TVB}_2$ representations to $\mathrm{STVB}_2$ conform to a $\Phi$-type extension. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:27:00 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nasser",
"Mohamad N.",
""
],
[
"Keshari",
"Vaibhav",
""
],
[
"Prabhakar",
"Madeti",
""
]
] |
2506.03807 | Chadha Degachi Ms | Chadha Degachi and Samuel Kernan Freire and Evangelos Niforatos and Gerd Kortuem | Understanding Mental Models of Generative Conversational Search and The Effect of Interface Transparency | Work in Progress | null | null | null | cs.HC cs.IR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The experience and adoption of conversational search is tied to the accuracy and completeness of users' mental models -- their internal frameworks for understanding and predicting system behaviour. Thus, understanding these models can reveal areas for design interventions. Transparency is one such intervention which can improve system interpretability and enable mental model alignment. While past research has explored mental models of search engines, those of generative conversational search remain underexplored, even while the popularity of these systems soars. To address this, we conducted a study with 16 participants, who performed 4 search tasks using 4 conversational interfaces of varying transparency levels. Our analysis revealed that most user mental models were too abstract to support users in explaining individual search instances. These results suggest that 1) mental models may pose a barrier to appropriate trust in conversational search, and 2) hybrid web-conversational search is a promising novel direction for future search interface design. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:27:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Degachi",
"Chadha",
""
],
[
"Freire",
"Samuel Kernan",
""
],
[
"Niforatos",
"Evangelos",
""
],
[
"Kortuem",
"Gerd",
""
]
] |
2506.03808 | Alice Lixuan Xu | Alice Lixuan Xu, Jorge S\'anchez Canales, Chiara Fusar Bassini, Lynn H. Kaack, Lion Hirth | Market power abuse in wholesale electricity markets | null | null | null | null | econ.GN q-fin.EC | http://creativecommons.org/licenses/by/4.0/ | In wholesale electricity markets, prices fluctuate widely from hour to hour and electricity generators price-hedge their output using longer-term contracts, such as monthly base futures. Consequently, the incentives they face to drive up the power prices by reducing supply has a high hourly specificity, and because of hedging, they regularly also face an incentive to depress prices by inflating supply. In this study, we explain the dynamics between hedging and market power abuse in wholesale electricity markets and use this framework to identify market power abuse in real markets. We estimate the hourly economic incentives to deviate from competitive behavior and examine the empirical association between such incentives and observed generation patterns. Exploiting hourly variation also controls for potential estimation bias that do not correlate with economic incentives at the hourly level, such as unobserved cost factors. Using data of individual generation units in Germany in a six-year period 2019-2024, we find that in hours where it is more profitable to inflate prices, companies indeed tend to withhold capacity. We find that the probability of a generation unit being withheld increases by about 1 % per euro increase in the net profit from withholding one megawatt of capacity. The opposite is also true for hours in which companies benefit financially from lower prices, where we find units being more likely to be pushed into the market by 0.3 % per euro increase in the net profit from capacity push-in. We interpret the result as empirical evidence of systematic market power abuse. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:28:27 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Xu",
"Alice Lixuan",
""
],
[
"Canales",
"Jorge Sánchez",
""
],
[
"Bassini",
"Chiara Fusar",
""
],
[
"Kaack",
"Lynn H.",
""
],
[
"Hirth",
"Lion",
""
]
] |
2506.03809 | Alexey Akimov | P.A. Aksentsev, V.A. Khlebnikov, I.S. Cojocaru, A.E. Rudnev, I.A. Pyrkh, D.A. Kumpilov, P.V. Trofimova, A.M. Ibrahimov, O.I. Blokhin, K.O. Frolov, S.A. Kuzmin, A.K. Zykova, D.A. Pershin, V.V. Tsyganok, A.V. Akimov | Phase stabilization and phase tuning of an optical lattice with a variable period | null | null | null | null | cond-mat.quant-gas physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Optical lattices play a significant role in the field of cold atom physics, particularly in quantum simulations. Varying the lattice period is often a useful feature, but it presents the challenge of maintaining lattice phase stability in both stationary and varying-period regimes. Here, we report the realization of a feedback loop for a tunable optical lattice. Our scheme employs a CCD camera, a computer, and a piezoelectric actuator mounted on a mirror. Using this setup, we significantly improved the long-term stability of an optical lattice over durations exceeding 10 seconds. More importantly, we demonstrated a rapid change in the optical lattice period without any loss of phase. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:31:05 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Aksentsev",
"P. A.",
""
],
[
"Khlebnikov",
"V. A.",
""
],
[
"Cojocaru",
"I. S.",
""
],
[
"Rudnev",
"A. E.",
""
],
[
"Pyrkh",
"I. A.",
""
],
[
"Kumpilov",
"D. A.",
""
],
[
"Trofimova",
"P. V.",
""
],
[
"Ibrahimov",
"A. M.",
""
],
[
"Blokhin",
"O. I.",
""
],
[
"Frolov",
"K. O.",
""
],
[
"Kuzmin",
"S. A.",
""
],
[
"Zykova",
"A. K.",
""
],
[
"Pershin",
"D. A.",
""
],
[
"Tsyganok",
"V. V.",
""
],
[
"Akimov",
"A. V.",
""
]
] |
2506.03810 | Shin-Ted Lin | Chang-Hao Fang, Shin-Ted Lin, Shu-Kui Liu, Henry Tsz-King Wong, Hao-Yang Xing, Li-Tao Yang, Qian Yue and Jing-Jun Zhu | Molecular Dynamics Simulations on Nuclear Recoils in Silicon Crystals towards Single Electron-Hole Pair Ionization Yields | 7 pages, 5 figures | null | null | null | astro-ph.IM cond-mat.mtrl-sci | http://creativecommons.org/publicdomain/zero/1.0/ | We have developed a novel methodology utilizing molecular dynamics simulations to evaluate the ionization yields of nuclear recoils in crystalline silicon. This approach enables analytical exploration of atomic-scale transport within the lattice without necessitating parameterization. The quenching factors across the nuclear recoil energy range from 20 eV to 10 keV have been thoroughly investigated. A remarkable agreement with experimental data is achieved, particularly for the minimal energy regime conducted to date, reaching the level of a single electron-hole pair. This work presents evidence of a crucial and fundamental distribution of the quenching factor, which can be associated to the collisional interactions underlying the transport phenomena. The region below 4 keV of the quenching factor, where discrepancies have been observed with the Lindhard's model, is found to be significantly attributed to the lattice binding effect and the specific crystal structure. In contrast, a gradual functional relationship is identified below approximately 100 eV, indicating that the quenching factor is influenced by the crystallographic orientation of the target material. From a distributional perspective, our analysis allows for the determination of the minimum exclusion mass for the dark matter nucleon elastic scattering channel at 0.29 $\mathrm{GeV}/c^2$, thereby significantly enhancing sensitivity for the sub-$\mathrm{GeV}/c^2$ mass region. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:31:51 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fang",
"Chang-Hao",
""
],
[
"Lin",
"Shin-Ted",
""
],
[
"Liu",
"Shu-Kui",
""
],
[
"Wong",
"Henry Tsz-King",
""
],
[
"Xing",
"Hao-Yang",
""
],
[
"Yang",
"Li-Tao",
""
],
[
"Yue",
"Qian",
""
],
[
"Zhu",
"Jing-Jun",
""
]
] |
2506.03811 | Sandra G\'omez | Mar\'ia Mallo, Susana G\'omez-Carrasco and Sandra G\'omez | Delayed photoisomerisation of the trans-PSB3 retinal toy model using on-the-fly quantum dynamics | null | null | null | null | physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | We explore the trans-cis photoisomerisation process in a representative retinal protonated Schiff base known as trans-PSB3, employing the quantum dynamics method direct dynamics variational multiconfigurational gaussian -- DD-vMCG -- in full dimensionality, i.e., 36 degrees of freedom on potential energy surfaces computed on-the-fly using the SA(2)-CAS(6,6)SCF electronic structure method with the 6-31G basis set. Although the toy molecule has been extensively studied using trajectory methods such as Tully Surface Hopping and Ab Initio Multiple Spawning, the on-the-fly quantum dynamics method DD-vMCG shows a trans-cis isomerisation hundreds of femtoseconds slower using the same electronic structure method, which can be explained in terms of the accesibility to the conical intersection connecting the ground and the excited state. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:32:50 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Mallo",
"María",
""
],
[
"Gómez-Carrasco",
"Susana",
""
],
[
"Gómez",
"Sandra",
""
]
] |
2506.03812 | Mahesh Nagargoje Dr. | Mahesh S. Nagargoje, Virginia Fregona, Giulia Luraghi, Francesco Migliavacca, Demitria A Poulos, Bryan C Good, and Jose Felix Rodriguez Matas | Impact of friction force and retrieval speed on in silico mechanical thrombectomies: a sensitivity analysis | null | null | null | null | physics.med-ph cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | Background: Mechanical Thrombectomy (MT) is a widely accepted first-line treatment for Acute Ischemic Stroke (AIS) and it has been studied using in vitro and in silico models. Thrombectomy outcomes have been performed for patient-specific cases using in silico models. However, until now, in vivo friction coefficients for stent-vessel, stent-clot, and clot-vessel interactions are unknown, but in vitro experiments have been attempted with significant standard deviations. These interactions and friction coefficients have been considered an important aspect of thrombectomy success. Objectives: In the current study, we explored the influence of variation in friction forces for stent-vessel, stent-clot, and clot-vessel interactions using virtual mechanical thrombectomy (VMT). We have performed three simulations for each interaction and varied friction coefficients around the standard deviation observed in the past in vitro studies. Results: (i) clot-vessel friction: higher friction leads to clot fragmentation and VMT failure. (ii) stent-clot friction: it is susceptible to VMT outcomes, with lower values showing the slippage of the clot while higher values lead to fragmentation. (iii) stent-vessel friction: higher friction shows compression of the stent in curved vessels and dislodgment of clot from stent retriever (SR) due to its compression, which leads to VMT failure. (iv) retrieval speed (RS): higher RS (>30 mm/s) leads to significant stent compression and unrealistic behavior of the SR. Conclusions: Analysis of results proposes the necessity for calculating accurate friction factor values and their implementation into in silico models, due to their sensitivity towards thrombectomy outcomes. Such in silico models mimic in vivo thrombectomy more closely and can be used in mechanical thrombectomy planning, management, and decision-making. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:34:03 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Nagargoje",
"Mahesh S.",
""
],
[
"Fregona",
"Virginia",
""
],
[
"Luraghi",
"Giulia",
""
],
[
"Migliavacca",
"Francesco",
""
],
[
"Poulos",
"Demitria A",
""
],
[
"Good",
"Bryan C",
""
],
[
"Matas",
"Jose Felix Rodriguez",
""
]
] |
2506.03813 | Lili Chen | Lili Chen, Changyang She, Jingge Zhu and Jamie Evans | Graph Neural Networks for Resource Allocation in Multi-Channel Wireless Networks | null | null | null | null | cs.LG eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As the number of mobile devices continues to grow, interference has become a major bottleneck in improving data rates in wireless networks. Efficient joint channel and power allocation (JCPA) is crucial for managing interference. In this paper, we first propose an enhanced WMMSE (eWMMSE) algorithm to solve the JCPA problem in multi-channel wireless networks. To reduce the computational complexity of iterative optimization, we further introduce JCPGNN-M, a graph neural network-based solution that enables simultaneous multi-channel allocation for each user. We reformulate the problem as a Lagrangian function, which allows us to enforce the total power constraints systematically. Our solution involves combining this Lagrangian framework with GNNs and iteratively updating the Lagrange multipliers and resource allocation scheme. Unlike existing GNN-based methods that limit each user to a single channel, JCPGNN-M supports efficient spectrum reuse and scales well in dense network scenarios. Simulation results show that JCPGNN-M achieves better data rate compared to eWMMSE. Meanwhile, the inference time of JCPGNN-M is much lower than eWMMS, and it can generalize well to larger networks. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:34:05 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Chen",
"Lili",
""
],
[
"She",
"Changyang",
""
],
[
"Zhu",
"Jingge",
""
],
[
"Evans",
"Jamie",
""
]
] |
2506.03814 | Annika Huch | Pamela Fleischmann, Annika Huch, Melf Kammholz, Tore Ko{\ss} | Jumbled Scattered Factors | null | null | null | null | math.CO cs.FL | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this work, we combine the research on (absent) scattered factors with the one of jumbled words. For instance, $\mathtt{wolf}$ is an absent scattered factor of $\mathtt{cauliflower}$ but since $\mathtt{lfow}$, a jumbled (or abelian) version of $\mathtt{wolf}$, is a scattered factor, $\mathtt{wolf}$ occurs as a jumbled scattered factor in $\mathtt{cauliflower}$. A \emph{jumbled scattered factor} $u$ of a word $w$ is constructed by letters of $w$ with the only rule that the number of occurrences per letter in $u$ is smaller than or equal to the one in $w$. We proceed to partition and characterise the set of jumbled scattered factors by the number of jumbled letters and use the latter as a measure. For this new class of words, we relate the folklore longest common subsequence (scattered factor) to the number of required jumbles. Further, we investigate the smallest possible number of jumbles alongside the jumbled scattered factor relation as well as Simon's congruence from the point of view of jumbled scattered factors and jumbled universality. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:37:22 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Fleischmann",
"Pamela",
""
],
[
"Huch",
"Annika",
""
],
[
"Kammholz",
"Melf",
""
],
[
"Koß",
"Tore",
""
]
] |
2506.03815 | Xu He | Tian Bai and Dianpeng Wang and Kuangqi Chen and Xu He | Adaptive Grid Designs for Classifying Monotonic Binary Deterministic Computer Simulations | 36 pages, 13 figures, 2 tables, 3 algorithms | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This research is motivated by the need for effective classification in ice-breaking dynamic simulations, aimed at determining the conditions under which an underwater vehicle will break through the ice. This simulation is extremely time-consuming and yields deterministic, binary, and monotonic outcomes. Detecting the critical edge between the negative-outcome and positive-outcome regions with minimal simulation runs necessitates an efficient experimental design for selecting input values. In this paper, we derive lower bounds on the number of functional evaluations needed to ensure a certain level of classification accuracy for arbitrary static and adaptive designs. We also propose a new class of adaptive designs called adaptive grid designs, which are sequences of grids with increasing resolution such that lower resolution grids are proper subsets of higher resolution grids. By prioritizing simulation runs at lower resolution points and skipping redundant runs, adaptive grid designs require the same order of magnitude of runs as the best possible adaptive design, which is an order of magnitude fewer than the best possible static design. Numerical results across test functions, the road crash simulation and the ice-breaking simulation validate the superiority of adaptive grid designs. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:38:00 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Bai",
"Tian",
""
],
[
"Wang",
"Dianpeng",
""
],
[
"Chen",
"Kuangqi",
""
],
[
"He",
"Xu",
""
]
] |
2506.03816 | Michel Destrade | Zhaoyi Zhang, Guo-Yang Li, Yuxuan Jiang, Yang Zheng, Artur L. Gower, Michel Destrade, Yanping Cao | Non-invasive measurement of local stress inside soft materials with programmed shear waves | null | Science Advances. Vol.9 (2023) eadd4082 | 10.1126/sciadv.add4082 | null | cond-mat.soft physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | Mechanical stresses in soft materials across different length scales play a fundamental role in understanding the function of biological systems and in the use of artificial materials for engineering soft machines and biomedical devices. Yet it remains a great challenge to probe local mechanical stresses in situ in a non-invasive, non-destructive manner, in particular when the mechanical properties are unknown. To address this challenge, we propose an acoustoelastic imaging-based method to infer the local mechanical stresses in soft materials by measuring the speed of shear waves induced by custom-programmed acoustic radiation force. Using a medical ultrasound transducer to excite and track the shear waves remotely, we demonstrate the application of the method by imaging uniaxial stress and bending stress in an isotropic hydrogel, and the passive uniaxial stress in a skeletal muscle. These measurements were all done without the knowledge of the constitutive parameters of the materials. These examples indicate that our method will find broad applications, ranging from health monitoring of soft structures and machines, to the diagnosis of diseases that alter stresses in soft tissues. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:39:49 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Zhang",
"Zhaoyi",
""
],
[
"Li",
"Guo-Yang",
""
],
[
"Jiang",
"Yuxuan",
""
],
[
"Zheng",
"Yang",
""
],
[
"Gower",
"Artur L.",
""
],
[
"Destrade",
"Michel",
""
],
[
"Cao",
"Yanping",
""
]
] |
2506.03817 | Julius Gonsior | Julius Gonsior, Tim Rie{\ss}, Anja Reusch, Claudio Hartmann, Maik Thiele, Wolfgang Lehner | Survey of Active Learning Hyperparameters: Insights from a Large-Scale Experimental Grid | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Annotating data is a time-consuming and costly task, but it is inherently required for supervised machine learning. Active Learning (AL) is an established method that minimizes human labeling effort by iteratively selecting the most informative unlabeled samples for expert annotation, thereby improving the overall classification performance. Even though AL has been known for decades, AL is still rarely used in real-world applications. As indicated in the two community web surveys among the NLP community about AL, two main reasons continue to hold practitioners back from using AL: first, the complexity of setting AL up, and second, a lack of trust in its effectiveness. We hypothesize that both reasons share the same culprit: the large hyperparameter space of AL. This mostly unexplored hyperparameter space often leads to misleading and irreproducible AL experiment results. In this study, we first compiled a large hyperparameter grid of over 4.6 million hyperparameter combinations, second, recorded the performance of all combinations in the so-far biggest conducted AL study, and third, analyzed the impact of each hyperparameter in the experiment results. In the end, we give recommendations about the influence of each hyperparameter, demonstrate the surprising influence of the concrete AL strategy implementation, and outline an experimental study design for reproducible AL experiments with minimal computational effort, thus contributing to more reproducible and trustworthy AL research in the future. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:41:37 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Gonsior",
"Julius",
""
],
[
"Rieß",
"Tim",
""
],
[
"Reusch",
"Anja",
""
],
[
"Hartmann",
"Claudio",
""
],
[
"Thiele",
"Maik",
""
],
[
"Lehner",
"Wolfgang",
""
]
] |
2506.03818 | Zhe Wang | Dong-Xu Yu, Ke-Qi Zeng, Zhe Wang | Understanding Flow Behaviors of Supercooled Liquids by Embodying Solid-Liquid Duality at Particle Level | 32 pages, 31 figures | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Supercooled liquids exhibit intricate flow behaviors, which progressively become nonlinear as flow rate increases. Conceptually, this complexity can be understood by the solid-liquid duality in Maxwell's understanding of materials' response to external load. Nevertheless, the microscopic foundation of this duality in supercooled liquids remains elusive, thereby impeding the modeling of flow behaviors from a microscopic viewpoint. The existence of dynamic heterogeneity adds to this difficulty. To tackle these problems, we propose the concept of local configurational relaxation time $\tau_\rm{LC}$, which is defined at the particle level. The spatial distribution of $\tau_\rm{LC}$ in flow is heterogeneous. Depending on the comparison between the local mobility measured by $\tau_\rm{LC}$ and the external shear rate, the response of local regions is either solid-like or liquid-like. In this way, $\tau_\rm{LC}$ plays a role similar to the Maxwell time. By applying this microscopic solid-liquid duality to different conditions of shear flow, we describe the emergence of shear thinning in steady shear, and predict the major characteristics of the transient response to start-up shear. Furthermore, we reveal a clear structural foundation for $\tau_\rm{LC}$ and the solid-liquid duality associated with it by introducing an order parameter extracted from local configuration. Thus, we establish a framework that connects microscopic structure, dynamics, local mechanical response, and flow behaviors for supercooled liquids. Finally, we rationalize our framework by leveraging the connection among structure, dynamics, and potential energy landscape (PEL). The PEL model illustrates how local structure, convection and thermal activation collectively determine $\tau_\rm{LC}$. Notably, it predicts two distinct response groups, which well correspond to the microscopic solid-liquid duality. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:41:43 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Yu",
"Dong-Xu",
""
],
[
"Zeng",
"Ke-Qi",
""
],
[
"Wang",
"Zhe",
""
]
] |
2506.03819 | Marc Aurel Vischer | Marc Aurel Vischer, Noelia Otero and Jackie Ma | Spatially Resolved Meteorological and Ancillary Data in Central Europe for Rainfall Streamflow Modeling | 6 pages, 1 figure | null | null | null | stat.ML cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We present a dataset for rainfall streamflow modeling that is fully spatially resolved with the aim of taking neural network-driven hydrological modeling beyond lumped catchments. To this end, we compiled data covering five river basins in central Europe: upper Danube, Elbe, Oder, Rhine, and Weser. The dataset contains meteorological forcings, as well as ancillary information on soil, rock, land cover, and orography. The data is harmonized to a regular 9km times 9km grid and contains daily values that span from October 1981 to September 2011. We also provide code to further combine our dataset with publicly available river discharge data for end-to-end rainfall streamflow modeling. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:44:33 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Vischer",
"Marc Aurel",
""
],
[
"Otero",
"Noelia",
""
],
[
"Ma",
"Jackie",
""
]
] |
2506.03820 | Sergiu Nisioi | Ahmad Mustapha Wali, Sergiu Nisioi | Automatic Correction of Writing Anomalies in Hausa Texts | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Hausa texts are often characterized by writing anomalies such as incorrect character substitutions and spacing errors, which sometimes hinder natural language processing (NLP) applications. This paper presents an approach to automatically correct the anomalies by finetuning transformer-based models. Using a corpus gathered from several public sources, we created a large-scale parallel dataset of over 450,000 noisy-clean Hausa sentence pairs by introducing synthetically generated noise, fine-tuned to mimic realistic writing errors. Moreover, we adapted several multilingual and African language-focused models, including M2M100, AfriTEVA, mBART, and Opus-MT variants for this correction task using SentencePiece tokenization. Our experimental results demonstrate significant increases in F1, BLEU and METEOR scores, as well as reductions in Character Error Rate (CER) and Word Error Rate (WER). This research provides a robust methodology, a publicly available dataset, and effective models to improve Hausa text quality, thereby advancing NLP capabilities for the language and offering transferable insights for other low-resource languages. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:46:19 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Wali",
"Ahmad Mustapha",
""
],
[
"Nisioi",
"Sergiu",
""
]
] |
2506.03821 | Michel Destrade | M. Destrade, Y. Du, J. Blackwell, N. Colgan, V. Balbi | Canceling the elastic Poynting effect with geometry | null | Physical Review E. Vol. 1007 (2023) L053001 | 10.1103/PhysRevE.107.L053001 | null | cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | The Poynting effect is a paragon of nonlinear soft matter mechanics. It is the tendency (found in all incompressible, isotropic, hyperelastic solids) exhibited by a soft block to expand vertically when sheared horizontally. It can be observed whenever the length of the cuboid is at least four times its thickness. Here we show that the Poynting effect can be easily reversed and the cuboid can shrink vertically, simply by reducing this aspect ratio. In principle, this discovery means that for a given solid, say one used as a seismic wave absorber under a building, an optimal ratio exists where vertical displacements and vibrations can be completely eliminated. Here we first recall the classical theoretical treatment of the positive Poynting effect, and then show experimentally how it can be reversed. Using Finite Element simulations, we then investigate how the effect can be suppressed. We find that cubes always provide a reverse Poynting effect, irrespective of their material properties (in the third-order theory of weakly nonlinear elasticity). | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:52:53 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Destrade",
"M.",
""
],
[
"Du",
"Y.",
""
],
[
"Blackwell",
"J.",
""
],
[
"Colgan",
"N.",
""
],
[
"Balbi",
"V.",
""
]
] |
2506.03822 | Fabian Karl | Fabian Karl, Ansgar Scherp | CRAWLDoc: A Dataset for Robust Ranking of Bibliographic Documents | Accepted at SCOLIA 2025 | null | null | null | cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Publication databases rely on accurate metadata extraction from diverse web sources, yet variations in web layouts and data formats present challenges for metadata providers. This paper introduces CRAWLDoc, a new method for contextual ranking of linked web documents. Starting with a publication's URL, such as a digital object identifier, CRAWLDoc retrieves the landing page and all linked web resources, including PDFs, ORCID profiles, and supplementary materials. It embeds these resources, along with anchor texts and the URLs, into a unified representation. For evaluating CRAWLDoc, we have created a new, manually labeled dataset of 600 publications from six top publishers in computer science. Our method CRAWLDoc demonstrates a robust and layout-independent ranking of relevant documents across publishers and data formats. It lays the foundation for improved metadata extraction from web documents with various layouts and formats. Our source code and dataset can be accessed at https://github.com/FKarl/CRAWLDoc. | [
{
"version": "v1",
"created": "Wed, 04 Jun 2025 10:52:55 GMT"
}
] | 2025-06-05T00:00:00 | [
[
"Karl",
"Fabian",
""
],
[
"Scherp",
"Ansgar",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.