Materials Science
Lattice dynamics of disordered phases of silicon and spatially resolved computations of thermal conductivity
A. Gautam, C. Ugwumadu, L. A. M. Rosset et al. (+3 more)
The
lattice dynamics of disordered systems (amorphous solids,
polymers and glasses) are often distinct from crystalline analogues. Examples include the "two-level systems" and the
Boson peak. Here, we show that even the venerable material amorphous silicon offers new wrinkles: (1) realistic networks studied with DFT-quality interatomic potentials show the existence of harmonic modes exhibiting anomalously large root-mean-square variation in atomic positions, and (2) these dynamics determine the processes of heat transport. We correlate the atomic dynamics with transport using the site-projected thermal conductivity [C. Ugwumadu et. al Phys. Status Solidi B 263, e202500316 (2026)], and
conjecture that these observations are relevant to other disordered systems.
Read the full abstract ▾
All authors A. Gautam, C. Ugwumadu, L. A. M. Rosset, R. M. Tutchton, S. Nakhmanson, D. A. Drabold
Artificial Intelligence
UniSkill: Learning Actor-Aligned Skill Proposals for an Evolving Policy
Yifei Lu, Cheng Liu, Dianzhi Yu et al. (+4 more)
Large
language model
agents can improve across tasks by retaining reusable skills distilled from prior interactions. Recent work jointly optimizes task execution and skill extraction, enabling the policy and skillbank to co-evolve. However, as the actor continues learning, rewarding skill proposals through their reuse in subsequent training steps may conflate skill benefits with actor improvement, while directly testing each proposed skill requires costly additional actor rollouts. In this paper, we introduce UniSkill, which uses a shared policy to interact with the environment and propose skillbank edits (Add, Update, or No Edit) from the resulting trajectories. Specifically, the actor learns from environment rewards, while contrastive action feedback guides skill proposal learning. This feedback provides an actor-
alignment signal by measuring how replacing the retrieved skill with a proposed skill changes the current actor's action log-likelihood gap between previously collected successful and failed trajectories from the same task, thereby avoiding new rollouts for each proposal. Since proposal-level feedback may suppress an otherwise appropriate edit operation when the proposed skill content scores poorly, we further apply skill-edit support regularization to preserve exploration. Empirically, UniSkill achieves strong performance, reaching 98.4% success on ALFWorld and 84.7% on WebShop while maintaining stable joint training. Further ALFWorld experiments show that UniSkill remains effective when the shared policy uses a smaller backbone. Our implementation is available at https://github.com/LimOkii/UniSKill.
Read the full abstract ▾
All authors Yifei Lu, Cheng Liu, Dianzhi Yu, Hui Xiang, Ji Zhang, Yuanchu Xiao, Rong Liang
Computer Vision
Beyond Anonymous Captions: Grounding Character Identity in Video Captioning and Question Answering
Anas Filali Razzouki, Killian Steunou, Khalil Guetari et al. (+3 more)
Linking people's appearance and actions to character identities is essential for understanding
video narratives. We present a framework for identity-aware
video captioning and person-centric question answering that combines automatic character identification, explicit spatial grounding, and task-specific adaptation. Starting from LSMDC v2 movie clips, our pipeline matches detected faces to actor reference images, tracks characters across frames, and builds inputs with identity-linked bounding boxes. A strong
vision-
language model generates identity-aware captions and questions, which are manually verified and filtered to create a
benchmark of 750 captioned clips and 3,000 person-centric questions. We study five grounding strategies combining textual coordinates with visual face or estimated person boxes across
Video-MLLM families at roughly 2B, 4B, and 8B parameters and larger frontier models. Combining visual face boxes with textual coordinates yields the most consistent performance across scales and significantly improves overall performance over coordinates alone. Smaller models tend to over-assign known identities when the queried person is not grounded, while larger models better recognize such UNIDENTIFIED cases. We introduce BAC by LoRA fine-tuning Qwen models at 2B, 4B, and 8B scales on about 32K identity-aware captioned clips. Across all scales, BAC outperforms every other evaluated model family of comparable size. BAC-8B reaches 93.20% overall QA accuracy, ranking behind only GPT-5.6 Sol among the frontier models evaluated in our study. Overall, explicitly communicating who is where, together with lightweight task-specific adaptation, substantially improves identity-aware
video understanding without changing the underlying architecture. We release the
benchmark, training data, code, and BAC checkpoints at https://github.com/momentslab/beyond-anonymous-captions.
Read the full abstract ▾
All authors Anas Filali Razzouki, Killian Steunou, Khalil Guetari, Thomas Kling, Mounîm El-Yacoubi, Yannis Tevissen
Probability
Limit shapes for centrally excited walks
Ahmed Bou-Rabee, Tuan-Minh Nguyen, Yuval Peres et al. (+1 more)
A centrally excited random walk on $\mathbb Z^d$ moves like simple random walk, except that its first step from each site other than the origin has a drift of fixed size toward it. Kozma (2007) conjectured that after $n$ steps the visited set approximates a ball with radius of order $n^{1/(d+1)}$. We prove this in every dimension $d\geq 2$ and show that the rescaled local times converge uniformly to a cone, which is the potential generated by the drift on a ball. The same argument shows that, when the drift is opposite to a subgradient of a norm, the visited set approximates the ball of that norm. In the plane, the outer and inner radii of the visited set differ by at most $n^{1/6}$ times a power of $\log n$, and the exponent $1/6$ cannot be lowered. Earlier versions of this work appeared in three independent drafts by Nguyen (arXiv:2610.07724), Bou-Rabee and Peres (arxiv:2610.10162v1) and Singh (arxiv:2610.12138). The present paper brings together the results and ideas from these drafts.
Read the full abstract ▾
All authors Ahmed Bou-Rabee, Tuan-Minh Nguyen, Yuval Peres, Arvind Singh
Instrumentation and Detectors
Wavelength-Resolved Timing of Near-Infrared Scintillation in Gaseous Xenon
Robert Hammann, Steffen Form, Humaira Nazneen et al. (+3 more)
The near-infrared (NIR) scintillation signal of xenon has the potential to enhance the performance of xenon-based detectors, but a robust understanding of its properties is a prerequisite for its application. In this work, we measure the NIR time response for $\alpha$-particle interactions in gaseous xenon in 12 wavelength bands spanning approximately $(950-1620)\,$nm. The time response is decomposed into a fast (ns-scale) component and a dominant slow ($\mu$s-scale) component. The fast-component spectrum coincides with strong atomic emission lines, predominantly 6p$\,\to\,$6s, whereas the slow-component spectrum follows the broad continuum centred around $1.3\,\mu$m. The pressure dependence of the continuum can be described by a two-step kinetic model in which the emitting state is formed by a three-body process and depopulated by a two-body collisional process, with density-independent terms in both rates. Interpreting the slower of these as radiative decay yields an intrinsic lifetime of $\sim3\,\mu$s. Theoretical calculations identify a neutral-excimer transition $0_u^- \rightarrow 0_g^-$ and an ionic-dimer transition $1/2_{u} \rightarrow 1/2_g$ as plausible candidates for the continuum emission, with the ionic-dimer transition showing better overall agreement with the measurements.
Read the full abstract ▾
All authors Robert Hammann, Steffen Form, Humaira Nazneen, Gabriel Schöpfer, Milan Ončák, Teresa Marrodán Undagoitia
Computer Vision
BagDINO: Multi-View Baggage Re-Identification with DINOv3
Vita Santa Barletta, Danilo Caivano, Rebecca Margiotta et al. (+2 more)
Mishandled checked baggage remains a recurrent issue in airport operations, and current recovery workflows still largely rely on tag-based tracking, which does not directly support visual identification when tag evidence is missing or unavailable. This paper investigates baggage re-identification as an instance-level
retrieval problem in a multi-camera setting, leveraging DINOv3 foundation-model representations to match a query image against a gallery of registered baggage images. A Torchreid-style BNNeck re-identification head is placed on top of a DINOv3 backbone, and parameter-efficient adaptation is performed via LoRA. Experiments are conducted on the MVB
benchmark using a progressive study that compares a fully frozen backbone against LoRA and fine-tuning strategies. Results indicate that parameter-efficient adaptation of foundation-model features provides an effective and stable approach for multi-view baggage re-identification under limited training data.
Read the full abstract ▾
All authors Vita Santa Barletta, Danilo Caivano, Rebecca Margiotta, Massimiliano Morga, Davide Pio Posa
Mesoscale and Nanoscale Physics
Beyond Equilibrium Raman Thermometry: Fermi-Level Control of the Anti-Stokes Response in Graphene
Martin Jindra, Luka Pirker, Tim Verhagen et al. (+6 more)
Non-
equilibrium electron and
phonon populations govern heat generation and dissipation in nanoscale devices, yet their direct characterization remains challenging. Anti-Stokes Raman scattering provides a sensitive probe of
phonon populations and local temperatures, although its interpretation becomes more complex when electronic and phononic populations are driven out of
equilibrium. Here, we investigate the gate-dependent anti-Stokes Raman response of monolayer
graphene and find that the anti-Stokes response deviates from that expected from an
equilibrium phonon population. Using a spatially localized microdroplet spectroelectrochemical platform, dielectric back-gates, and chemical doping, we observe a strong and reversible suppression of the G-mode anti-Stokes intensity as the
Fermi level is tuned away from the Dirac point, while the Stokes intensity remains essentially unchanged. By contrast, the G-mode frequency and linewidth of both the Stokes and anti-Stokes branches evolve with doping as expected from the established Kohn-anomaly picture. The anti-Stokes/Stokes intensity ratio decreases by several-fold over the measured range and displays a pronounced, gate-dependent power response with maximum sensitivity at charge neutrality. Comparison with previous studies of non-
equilibrium carrier dynamics in doped
graphene suggests an important role of
Fermi-level-dependent electronic phase space, which can influence both hot-carrier-driven
phonon populations and electronically mediated Stokes-anti-Stokes scattering. Our findings establish anti-Stokes Raman scattering as a sensitive probe of electron-
phonon coupling and non-
equilibrium phonon dynamics in
graphene.
Read the full abstract ▾
All authors Martin Jindra, Luka Pirker, Tim Verhagen, Ciarán Fowley, Martin Kalbáč, Matěj Velický, Riichiro Saito, Ado Jorio, Otakar Frank
Analysis of PDEs
Sharp noncommutative Hilbert and Riesz inequalities: classification and spectral corrections
Zhouyu Long
For every $1<p<\infty$, with $p^*=\max\{p,p/(p-1)\}$, we determine the sharp
Hilbert norm $C_p=\cot(\pi/(2p^*))$ on $L^p(\mathbb{R};\mathcal{S}_p)$ and the sharp uniform bound $D_p=\csc(\pi/p)$ for retained and strict Riesz projections of maximal subdiagonal
algebras in finite tracial von Neumann
algebras. The
Hilbert value is also the common
Fourier–Schur completely bounded norm. For finite
matrices, a nonnegative logarithmic determinant defect connects Heinävaara's tracial joint spectral measure to
eigenvalues. The resulting transfer identity classifies finite signed real ridge inequalities on nilpotents and identifies their least additive
eigenvalue corrections. A dual
theorem characterizes weak-* limits of weighted angular measures of finite nilpotents. In general finite tracial
algebras, separate positive-kernel and sharp determinant-covariance comparisons yield the Riesz bounds. Consequences include the sharp Volterra constant, Riesz–Titchmarsh and shifted-transform formulas, joint real-pencil norms, and exact triangular and operator-valued
Fourier Riesz norms.
Read the full abstract ▾
Rings and Algebras
On computing deformations of contraction algebras of type E
Agata Smoktunowicz and Alicja Smoktunowicz
We construct deformations of contraction
algebras of type E to semisimple
algebras. We consider the class given by the potential x^3+xy^3+p(y), where p(y) is a non-zero
polynomial of order larger or equal to 4.
Computer Vision
HeiCo-FOCUS: A Clinically Grounded Dataset for Long-Context Video Understanding
Leon Mayer, Lucas Luttner, Patrick Godau et al. (+42 more)
Recent advances in
Vision-
Language Models (VLMs) have led to rapid progress in
video understanding across a wide range of
benchmark tasks. However, existing evaluations largely focus on short-term
reasoning, failing to assess a critical capability: maintaining cumulative temporal consistency over extended time horizons. To close this evaluation gap, we introduce HeiCo-FOCUS, a clinically grounded
dataset for evaluating long-context
video understanding through the task of Foreign Object Contextual Understanding in Surgery. Built on a
dataset of Heidelberg Colorectal surgeries, this task requires models to continuously track multiple objects as they are inserted, manipulated, occluded, and removed over procedures lasting up to hours. HeiCo-FOCUS comprises 30,000 visual question answering (VQA) pairs covering five core capabilities: object recognition, temporal grounding, aggregation, event and procedural understanding, and complex
reasoning. The
dataset was constructed through a rigorous multi-stage annotation pipeline involving large-scale crowd annotation and 39
surgical domain experts to ensure high quality and clinical relevance. To systematically probe model behavior, we introduce a multi-track evaluation framework that progressively increases temporal and contextual demands from single frames to full procedures. Experiments with ten frontier VLMs show that HeiCo-FOCUS tasks are far from solved: only around half of the models clearly outperform a text-only baseline. Across the
video tracks, models perform best on event and procedural understanding (mean Accuracy: 56.5% across all models), while temporal grounding remains particularly challenging for all evaluated models (mean Accuracy: 19.7%). We therefore expect HeiCo-FOCUS to serve as a catalyst for the development of models capable of reliable, temporally consistent
reasoning over hours-long
videos.
Read the full abstract ▾
All authors Leon Mayer, Lucas Luttner, Patrick Godau, Kai Fritzsche, Annika Reinke, Leonie Boland, Jule Brandt, Janne Heinecke, Chloe K. Nobuhara, Niklas Holzwarth, Evangelia Christodoulou, Marcel Knopp, Dominik Michael, Pascale Piermarco, Saliq Neyaz, Korhan Derin Özarslan, Jakob Hennighausen, Carlos Aumente-Maestro, Tim Rädsch, Dheeraj Baji, Peter Maximilian Full, Finn Aichholz, Justus Veit Erpenbeck, Linus Finn Schott, Bastian Winkelhausen, Claas de Boer, Bianca Güttner, Anneli Hummel, Gregor Just, Max Kirchner, Chenyang Li, Rozenn Raffaut, Ariel Rodriguez, Danush Kumar Venkatesh, Kevin Wang, Jinjing Xu, Mona Sheikh Zeinoddin, Salman Khan, Thomas M. Pausch, Stefanie Speidel, Danail Stoyanov, Daniel A. Hashimoto, Fiona R. Kolbinger, Thomas G. Weiser, Lena Maier-Hein
Signal Processing
Near-Field Beamforming Design for Multi-IRS Wireless Beam Routing
Tao Wang, Changsheng You and Changchuan Yin
This letter investigates efficient near-field beamforming designs for wireless communication systems aided by multiple extremely large-scale intelligent reflecting surfaces (Multi-XL-IRSs). Previous studies have mostly assumed far-field channel conditions, based on which angle-based beamforming is adopted for the base station (BS) and IRSs. However, this approach may lead to degraded performance in practical near-field scenarios. To address this issue, we consider near-field channel modeling and formulate an optimization problem aimed at maximizing the received power at the user equipment (UE) by jointly optimizing the beamforming at the BS and IRSs. Given the intractability under the beam training (BT)-based communication framework, the problem is revisited and reformulated. Subsequently, an efficient
algorithm based on alternating optimization (AO) is proposed to solve the reformulated problem. Numerical results demonstrate that, in typical setups, our AO-based beamforming designs provide over 170% and 20% received power gains over conventional angle-based beamforming and beam focusing, respectively. Furthermore, the power scaling law derived under far-field conditions no longer holds in near-field scenarios.
Read the full abstract ▾
Computation and Language
HySPE: Positional Encoding via Symplectic Dual Shears
Zhongping Ji
We introduce
Hyperbolic Symplectic Positional Encoding (HySPE), grounding positional
attention in non-compact
symplectic transformations. While canonical Rotary Position
Embedding (RoPE) parameterizes the compact,
elliptic branch of $\Sp(2,\R)$ via rotations, HySPE operationalizes its
hyperbolic branch via a damped symmetric composition of dual shears, yielding a conformally
symplectic contraction with two spectral decay rates per channel pair. To eliminate the exponential representation drift inherent to naive absolute factorizations, we diagonalize the operator in its invariant eigenbasis and introduce blockwise coordinate rebasing with adaptive centered execution. This guarantees length-independent numerical bounds while matching cached RoPE forward latency (7.21\,ms on an RTX 4090). On TinyShakespeare, HySPE-UltraLong maintains an invariant perplexity of 4.810 up to $16\times$ zero-shot extrapolation ($L=4096$), whereas RoPE degrades to 131.198. Scaled to a 51M-parameter subword
Transformer on WikiText-103 ($L_{\text{train}}=512$), HySPE closely matches RoPE in-domain while robustly extrapolating to length 8192, reducing tail perplexity by 83.9% over RoPE. While these controlled experiments establish HySPE's extrapolation robustness and numerical stability, evaluating its scaling behavior on large-scale foundation models remains an important direction for future investigation.
Read the full abstract ▾
Data Structures and Algorithms
3SUM Is Really Hard: A Real-to-Integer Reduction
Nick Fischer, Adam Polak and Jonas Schmidt
We show that if the 3SUM problem on integer-valued inputs can be solved in truly subquadratic time, then it can also be solved in truly subquadratic time on real-valued inputs. This answers an open problem posed by Chan, Vassilevska Williams, and Xu [STOC 2022], and constitutes the first such tight real-to-integer self-reduction in fine-grained complexity. Our proof relies on a surprising combination of the Frank–Tardos
theorem from optimization with Freiman-type
theorems from additive combinatorics.
Read the full abstract ▾
Metric Geometry
The centered logarithmic Brunn–Minkowski inequality
Yude Liu, Qiang Sun and Ge Xiong
The logarithmic Brunn–Minkowski and logarithmic Minkowski inequalities are proved for convex bodies whose centroids are at the origin, with equality precisely for independent positive dilations of common direct summands. This characterizes when centered convex bodies have the same cone-volume measure. The \(L_p\) Brunn–Minkowski and Minkowski inequalities are also obtained for every \(p>0\), with equality only for positive dilates, and the corresponding uniqueness of centered \(L_p\) surface area measures is established.
Read the full abstract ▾
Symplectic Geometry
Integration and quantization of 1-shifted coisotropics
Yassine Ait Mohamed and Maxence Mayrand
To quantize the reduction of a coisotropic submanifold of a Poisson
manifold, Cattaneo and Felder constructed a homotopy Poisson
algebra together with a deformation quantization. Its degree-zero
cohomology is the reduced Poisson
algebra, while the full homotopy structure retains information about the formal
embedding even when the reduced space is singular or trivial. On the integration side, Cattaneo showed that a coisotropic submanifold integrates to a Lagrangian subgroupoid of any
symplectic groupoid integrating the ambient Poisson
manifold. It is now recognized that both constructions fit into shifted
symplectic geometry: the Lagrangian subgroupoid presents a $1$-shifted Lagrangian morphism of differentiable stacks, and the homotopy Poisson
algebra describes the associated $0$-shifted Poisson structure. However, such morphisms form only a special class of $1$-shifted Lagrangians. We extend this picture to $1$-shifted Lagrangians and, more generally, $1$-shifted coisotropics in $1$-shifted
symplectic differentiable stacks. We establish an integration
theorem for their infinitesimal description in terms of twisted Dirac structures and construct a natural Poisson bracket on invariant functions. Under regularity hypotheses, we equip the associated Lie algebroid complex with a flat $P_\infty$-structure and a curved $A_\infty$-quantization, independent of the auxiliary choices up to isomorphism. The induced bracket in degree zero recovers the reduced Poisson bracket, and suitable cohomological vanishing conditions yield a deformation quantization of invariant functions. This extends the Cattaneo–Felder picture to more general reduction procedures, with the same infinitesimal data governing integration and quantization.
Read the full abstract ▾
Cryptography and Security
On the Reliability of LLM-Based Vulnerability Patching Benchmarks
Dang K Le, Wenxuan Shi and Xinyu Xing
Large
language models (LLMs) have shown strong potential for automated vulnerability patching, but current
benchmarks can substantially distort reported performance. Drawing on extensive experience developing, running, and stress-testing such frameworks, we identify under-examined pitfalls across three dimensions: (1)
agent-level factors, where prompting, tool availability, and detailed instructions can raise success rates without improving developer-aligned patch quality; (2) framework-level factors, where permission errors, infrastructure bugs, and timeout handling can silently suppress or inflate performance; and (3)
dataset-level factors, where bug reports and single proof-of-concept (PoC) tests fail to capture whether patches address root causes or follow developer intent. We curate 112 historical bugs from 84 open-source C/C++, Go, and Rust projects, each with PoC tests, regression tests, and additional developer tests that assess
alignment with the original developers' design principles. Through controlled experiments and case studies, we show that LLMs can achieve high PoC passing rates under ideal conditions, yet
benchmark execution choices can materially change measured success. More importantly, developer-test passing rates remain low and improve only marginally with newer models, suggesting that models increasingly suppress symptoms without consistently producing upstream-quality fixes. These results show that
benchmark scores are highly sensitive to evaluation design, and we provide practical guidelines for more rigorous, reliable, and reproducible evaluation.
Read the full abstract ▾
Cryptography and Security
Pump-and-Dump meets Honeypot Tokens: Detection and Analysis of Telegram Bait-and-Trap Schemes
Federico Cernera, Massimo La Morgia, Alessandro Mei et al. (+1 more)
While Pump-and-Dump schemes have been extensively studied on centralized exchanges (CEXs), how they operate in decentralized exchanges (DEXs) remains largely unexplored. We monitor 83 Telegram channels used to coordinate Pump-and-Dump campaigns and collect 3,677 events across the BNB Smart Chain and Ethereum. Our analysis reveals that, although these operations superficially resemble CEX-based Pump-and-Dump schemes, their underlying mechanism is fundamentally different. Rather than manipulating prices, organizers orchestrate deceptive Pump-and-Dump campaigns around honeypot
tokens whose smart contracts allow Telegram subscribers to buy but prevent them from selling. Unaware of this restriction, victims purchase these
tokens and are unable to recover their funds. We refer to this new fraud mechanism as Bait-and-Trap. We show that Bait-and-Trap provides a substantially more reliable profit strategy than Pump-and-Dump schemes: organizers gain in 99.3% of cases, extracting over $7 million in profit. To counter this threat, we develop a transaction-simulation tool that detects honeypot
tokens before purchase by testing whether a user can successfully sell against the
token's live contract state, providing a practical defense against Bait-and-Trap operations.
Read the full abstract ▾
All authors Federico Cernera, Massimo La Morgia, Alessandro Mei, Francesco Sassi
Distributed Computing
Reconciling Bottom-Up Metrics with Top-Down Reporting for Cloud Carbon Accounting
Philipp Wiesner, Loïc Lannelongue, Alexander Acker et al. (+1 more)
Organizations seeking to reduce the carbon footprint of their cloud applications must rely on two largely disconnected ways of measuring it, each with important limitations. Bottom-up metrics like Software Carbon Intensity (SCI) provide signals for carbon-aware optimization, but tenants lack the information needed to account for many provider-side overheads. Top-down reports by cloud providers provide a more comprehensive view of a tenant's carbon footprint but are coarse and methodologically opaque. Because the two approaches differ in scope and reporting frequency, the current state of the art treats them as decoupled: bottom-up metrics for optimization, top-down reports for corporate disclosure. We argue that the two signals should be reconcilable. A per-workload metric whose improvements never surface in the provider's audited report will not be adopted for accountability at scale. We survey the state of the art in cloud carbon accounting and propose a
vision for reconciled SCI (rSCI): a per-workload metric that re-anchors a bottom-up energy estimate to the cloud provider's top-down report through a residual. By decomposing and allocating this residual according to its physical drivers, rSCI can preserve operational incentives while also attributing idle capacity and
embodied carbon to the workloads that drive them. We show that today's top-down reporting is still too coarse and methodologically inconsistent to support credible reconciliation and lay out concrete steps in provider reporting and standards to make it practical.
Read the full abstract ▾
All authors Philipp Wiesner, Loïc Lannelongue, Alexander Acker, Odej Kao
Algebraic Geometry
Counterexamples to the Polishchuk–Van den Bergh conjecture on curves
Shengyong Pan
We construct effective finite group actions on smooth projective complex curves for which
Conjecture A of Polishchuk and Van den Bergh fails. For every $h\geq1$, an explicit $S_3$ action on the smooth projective model of $y^2=t^{12h-5}+t$ has quotient of genus $h$ and exactly one ramification orbit, with inertia of order three. The fixed quotient for the three-cycle class consists of two points, but its derived category admits no $\C$-linear exact fully faithful
embedding into the equivariant derived category. More generally, when the coarse quotient has positive genus, we prove that the maximum number of pairwise completely orthogonal exceptional objects is $\sum_i\lfloor e_i/2\rfloor$, where $e_i$ are the inertia orders. We also characterize conjugacy-class decompositions with
embeddings linear over the coarse quotient: such a decomposition exists if and only if $|[g]\cap H_i|\leq\lfloor e_i/2\rfloor$ for every nonidentity conjugacy class and every branch inertia group $H_i$. The sufficiency proof uses coloured paths of type $A$ and derived reflection functors, and produces kernels on the natural fibre products. The known ordered curve decomposition still yields a decomposition indexed by the connected components of the fixed quotients.
Read the full abstract ▾
Fluid Dynamics
Extension of a mathematical analogy between vortical flows without surface tension and irrotational capillary flows
Valentin D. Kunz, Saleh Tanveer and Darren G. Crowdy
In a recent paper, a mathematical analogy was found between two physically distinct classes of two-dimensional free boundary problems in inviscid fluid mechanics. One class involves free surface flows with constant vorticity in the bulk but without boundary surface tension, the other comprises free surface flows without vorticity in the bulk but with non-zero boundary surface tension (capillarity). Several examples were given of pairs of physically-distinct free boundary problems that amount to solving the same mathematical problem, albeit with the possibility of a single constraint arising for the capillary flows. The present work shows that this analogy extends to another pair of physically-distinct problems: a hollow
vortex equilibrium in a shear flow with a third order irrotational strain and a steady bubble with surface tension in a simple straining flow. For both problems, we construct an isolated exact solution when a parameter $\gamma=0$ and provide numerical solutions for non-zero $\gamma$. An interesting feature is that non-trivial conformal mapping identities emerge from the class of problems with capillarity that apply equally for the hollow
vortex problems when the parameters are appropriately constrained.
Read the full abstract ▾
Computation and Language
InterView-C: A Synchronized Multimodal Corpus of VR Avatar-Mediated Survey Interviews
Patrick Schrottenbacher, Leon Hammerla, Lydia Kleine et al. (+2 more)
We present InterView-C, a German
multimodal corpus of 27 survey interviews conducted entirely in virtual reality, with both interlocutors represented by avatars. The corpus aligns spoken interaction with synchronized behavioral data, including gaze, head and body movement, facial behavior, hand and finger tracking. Its reference transcripts and linguistic annotations provide a reliable interface between this
multimodal spoken interaction and predominantly text-based NLP methods. This interface is important because automatically transcribing
speech can distort linguistically relevant information, while downstream models trained on existing resources may additionally face transfer challenges when applied to transcribed spoken data. InterView-C therefore provides word-timed and manually post-edited verbatim transcripts for all 54 recordings, interview-item timings, questionnaire responses and negation cue and scope annotations for 1,422 sentences, 1,398 of them doubly annotated ({\alpha}=0.87 for cues; {\alpha}=0.81 for scopes). We demonstrate both challenges empirically: nine open-weight ASR systems disproportionately misrecognize short closed answers and number words, while negation models trained on existing corpora show lower and highly variable performance on our transcribed interviews than a model trained on the InterView-C annotations. InterView-C thus enables linguistic analyses of spoken interaction while retaining their
alignment with rich
multimodal behavior.
Read the full abstract ▾
All authors Patrick Schrottenbacher, Leon Hammerla, Lydia Kleine, Doris Stingl, Alexander Mehler
Quantum Algebra
Higher Fomin-Kirillov algebras are infinite-dimensional
N. Andruskiewitsch, I. Angiono, I. Heckenberger et al. (+3 more)
We show that the Fomin–Kirillov
algebras $FK_n$ are infinite-dimensional for $n\geq6$. This answers a question posed by Fomin and Kirillov in 1997. To prove this, we demonstrate that the Nichols
algebras supported on the conjugacy class of transpositions of the symmetric group are infinite-dimensional for both relevant 2-cocycles. For this, we develop a new method of partially semisimplifying a Nichols
algebra along an
automorphism of order $p$ in characteristic $p$.
Read the full abstract ▾
All authors N. Andruskiewitsch, I. Angiono, I. Heckenberger, S. Lentner, C. Reiher, L. Vendramin
Quantum Physics
Inequivalent Quantum Resources from Multipartite State Discrimination
Srijani Das, Tuhin Paul, Manasi Patra et al. (+2 more)
Any two orthogonal
quantum states can be perfectly distinguished via local operations and classical communication (LOCC), yet whether all orthogonal multipartite pairs incur an identical operational cost remains an open question. Here, we address the resource consumption required for local state discrimination across multipartite pair of states. We demonstrate that distinct pairs of orthogonal multipartite maximally entangled states require unequal amounts of
quantum resources to achieve local distinguishability. Extending our framework beyond individual pairs, we analyze larger sets of both locally distinguishable and indistinguishable multipartite orthogonal states, revealing a nontrivial hierarchy in their
entanglement consumption. These results show that the resource cost of LOCC discrimination is state-dependent, providing insight into the fine-grained structure of multipartite nonlocality.
Read the full abstract ▾
All authors Srijani Das, Tuhin Paul, Manasi Patra, Atanu Bhunia, Ramij Rahaman
Statistical Methodology
When Can You Ship on Evals Alone? Trial-Level Surrogacy for Offline Evaluation of LLM Systems
Mårten Schultzberg
Offline evals increasingly guide changes to LLM-powered products, but showing that an eval agrees with human judgment does not tell a team which changes it can ship without an A/B test. That decision depends on whether eval treatment effects predict online treatment effects across a class of interventions. We call this requirement trial-level surrogacy and show that it is logically independent of unit-level (Prentice) surrogacy. In historical launch logs, sampling noise in both effect estimates attenuates the observed relationship. Under a bivariate normal working model, subtracting the reported within-intervention variances recovers the between-intervention covariance, and the corrected moments give a closed-form ship/test/kill rule that bounds the probability of shipping a change with a non-positive online effect. LLM judges add a separate distortion. When the judge is equally accurate on both arms, eval effects are only rescaled, but arm-dependent accuracy can reverse their sign. Once deployed, the eval-to-outcome relationship can no longer be observed in the ship and kill regions, so drift there becomes undetectable without randomized audit experiments. On 183 interventions from the Upworthy Research Archive, the pooled eval-to-outcome relationship is weak despite precise measurement on both sides, and the illustrative gate sends every intervention to an A/B test.
Read the full abstract ▾
Software Engineering
AdaT$^2$: Adaptive Test Transformations for Black-Box Boundary Testing of Conversational Agents
Liting Lin, Boxi Yu, Qinghua Xu et al. (+3 more)
Conversational
agents based on large
language models (LLMs) must comply with policies. Each condition in a policy draws a boundary between user requests, and the
agent must behave differently on its two sides. We present AdaT$^2$, which extracts statements from the
agent's replies in exploratory conversations with an LLM acting as the user, and uses the statements to guide boundary test generation. Each statement describes one condition and the behavior expected when the condition holds. Besides plain tests guided by single statements, AdaT$^2$ writes transformed tests guided by pairs of a statement and a test transformation instruction, such as "omit one required input". The instruction of a pair can move a test to the other side of the statement's boundary or to another boundary. The statements and instructions form far more pairs than a run can try, and many pairs are not applicable. Adaptive pair selection therefore chooses the statement of each pair by novelty and the instruction with the bandit
algorithm Bayes-UCB, which learns from whether earlier pairs yielded a test and whether the
agent passed it according to an LLM judge. Our
benchmark counts the two sides of each boundary separately and distinguishes boundaries explicitly defined by the
agent's
prompt, tool code, or knowledge base from boundaries that the
agent's LLM infers from domain knowledge. On four domains of $\tau^3$-bench, 62.7% to 83.3% of AdaT$^2$'s tests are valid boundary tests whose expected behavior is explicitly defined, higher in every domain than AgentEval's (47.7% to 68.1%). Transformed tests add 13 to 46 explicitly defined boundaries that plain tests miss. As regression tests, AdaT$^2$'s test suites detect all eight seeded policy faults in the airline domain and four of eight in the retail domain, and AgentEval's test suites, with fewer than a third as many tests, detect five and two, respectively.
Read the full abstract ▾
All authors Liting Lin, Boxi Yu, Qinghua Xu, Yuzhong Zhang, Lionel Briand, Emir Muñoz
Other Condensed Matter
Progress and Prospect of AI in ARPES Workflow
Sandy Adhitia Ekahana, Aalok Tiwari, Pratik Saud et al. (+4 more)
Artificial intelligence (AI) is becoming an increasingly useful tool across the experimental sciences, including angle-resolved photoemission spectroscopy (ARPES), which routinely produces large, multidimensional
datasets of electronic structure. Recent advances in AI and machine learning (ML) have opened new opportunities across the entire ARPES workflow, from automated sample preparation and real-time data acquisition to post-experiment data analysis and comparison with theoretical calculations. Despite this progress, a comprehensive review of ML applications, their capabilities, and reliability across the different stages of ARPES workflow is still lacking. In this review, we first introduce ML methods that are most relevant to experimentalists working in condensed matter physics and materials science. We then follow the ARPES workflow, reviewing existing ML applications at each step and discussing their advantages, limitations and potential for future development. We also examine the current ARPES data landscape, where several open databases are available but remain relatively small and fragmented compared with large, shared
datasets such as ImageNet. Given these limitations, we suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES
dataset repository. Finally, we discuss our perspectives on the future of AI within the ARPES workflow using a six-level framework of laboratory automation, highlighting the opportunities and challenges in moving toward a fully autonomous, self-driving ARPES laboratory.
Read the full abstract ▾
All authors Sandy Adhitia Ekahana, Aalok Tiwari, Pratik Saud, Aaron Bostwick, Chris Jozwiak, Eli Rotenberg, Jyoti Katoch
Number Theory
Generating sets for maximal orders in rational quaternion algebras
Kirsten Eisenträger, Eyal Z. Goren, Annamaria Iezzi et al. (+6 more)
Given a maximal order $\mathfrak{O}$ in a rational definite quaternion
algebra, and a
prime $\ell$ coprime to the discriminant of $\mathfrak{O}$, this paper considers subsets of $\mathfrak{O}$ consisting of elements with $\ell$-power norms that together generate $\mathfrak{O}$ as a $\mathbb{Z}$-
algebra. We prove two
theorems about the existence of such sets: the first states that $\mathfrak{O}$ is generated by elements of norm $\ell^k$ for any $k$ larger than an explicit bound, and the second states that there is a generating set for $\mathfrak{O}$ consisting of at most three elements, each with norm a power of $\ell$. We discuss implications for the study of supersingular isogeny graphs. As steps towards these
theorems, for quaternion orders $\mathcal{O}$ that are not necessarily maximal, we also prove structural results about the
algebra of Brandt
matrices for $\mathcal{O}$ and explicit bounds on the coefficients of the theta function of $\mathcal{O}$. Computational experiments are also discussed.
Read the full abstract ▾
All authors Kirsten Eisenträger, Eyal Z. Goren, Annamaria Iezzi, Harun Kir, Eda K{\i}r{\i}ml{\i}, Jonathan R. Love, William E. Mahaney, Jennifer Park, Maria Sabitova
Computation and Language
LLM4Impact: Integrating Heterogeneous Information for Scientific Impact Prediction
Yong Cao, Markus Flicke, Haoyu He et al. (+2 more)
Predicting the future impact of a newly published paper is challenging because it must be inferred from heterogeneous evidence available at publication time. Existing approaches often rely on a single source of information or combine multiple sources without accounting for their different predictive roles. In this paper, we present LLM4Impact, an evidence-aware method for scientific impact prediction that learns to represent, integrate, and calibrate heterogeneous information. LLM4Impact combines semantic, graph, LLM, and temporal representations, and injects graph information into a frozen LLM through continuous prefix
tokens. A context aware gating mechanism adaptively weights different evidence, while a separate calibration module accounts for domain and temporal variation in citation scales. We further construct a large-scale
benchmark dataset with 2 million papers, leakage-safe point-in-time heterogeneous ego graphs, temporal splits, and both year-level and month-level citation targets. Experiments show that LLM4Impact consistently outperforms strong semantic, graph, and LLM based baselines, with a 10.13% reduction in year RMSE on the in distribution test set and a 6.87% reduction under out-of-domain distribution. Our results reveal that the value of such evidence is context dependent: different papers benefit from different sources, while domain and publication time affect how evidence translates into citations. This finding motivates adaptive evidence selection and context-conditioned calibration rather than simply richer representations. We will release our code,
benchmark, and an interactive web demonstration upon publication.
Read the full abstract ▾
All authors Yong Cao, Markus Flicke, Haoyu He, Katrin Renz, Andreas Geiger
Operating Systems
Behind a Simple Read: Understanding Work and Waiting in the Linux I/O Stack
Yang Shen, Wenzhe Zhang, Yong Dong et al. (+4 more)
A simple read interface provides uniform semantics but unpredictable performance. Using synchronous large-buffer Linux reads, we decompose the buffered-read path into work volume, processing time, and critical-path exposure. Reducing local cost or accelerating request submission need not shorten a complete read. Linux uses large folios to reduce metadata work and I/O-copy overlap to hide most cold-read copying; an experimental kernel prototype shows that more aggressive overlap and parallelism can shorten the host path when these mechanisms fail. Device waiting dominates SSD cold reads. We abstract first-completion wait F and subsequent completion capacity B_CQ from request-batch timelines. Real-SSD experiments characterize their variation with request and batch size, while MQSim traces explain their internal origins. A wait chain along the folio-bio-request mapping connects device completion to upper-layer waiting. Under strict batching, the median absolute cross-layer prediction error is at most 7.43%. Finally, we apply the completion model and wait chain to request supply, Linux readahead, data layout, and waiting policies. Cases cover gains, no gains, and reversals, comparing predictions with interventions to close the loop from observation and modeling to understanding and control and provide a measurable basis for layered I/O decisions.
Read the full abstract ▾
All authors Yang Shen, Wenzhe Zhang, Yong Dong, Min Xie, Wei Zhang, Huijun Wu, Zhenwei Wu
Materials Science
Surface Physics of Low-Dimensional Pnictogen Chalcohalides and Its Impact on Photovoltaic Technologies
Cibrán López, David Rovira, Álex Jiménez-Arguijo et al. (+3 more)
Low-dimensional pnictogen chalcohalides (MChX; M = Bi, Sb; Ch = S, Se; X = I, Br) are promising non-toxic
semiconductors for
solar energy conversion, yet the power conversion efficiencies of MChX
solar cells remain below 10%, far from their Shockley-Queisser limit of $\sim$30%. Efficient charge extraction depends critically on the band-edge
alignment between the absorber and the charge-selective contacts, which is a surface property. Here, we combine first-principles calculations with device-level modelling to investigate the surface energetics and band
alignments across the MChX family. An exhaustive sampling of
crystal orientations, including low-
symmetry facets, identifies the (011) and (010) surfaces as the most stable terminations in all eight compounds. Their formation energies differ by only $\approx 0.01$-$0.03\mathrm{J\,m^{-2}}$, yet their band edges are shifted rigidly with respect to each other by up to $0.8$eV. Drift-
diffusion simulations of BiSBr show that the coexistence of such facets can reduce the open-
circuit voltage by up to $0.6$V, even in an otherwise ideal, defect-free absorber. Surface orientation thus emerges as a hidden design parameter for charge extraction. This study may help to explain the gap between the theoretical and experimental efficiencies of MChX
solar cells, and it provides new design principles for MChX-based photovoltaic and photocatalytic technologies.
Read the full abstract ▾
All authors Cibrán López, David Rovira, Álex Jiménez-Arguijo, Zacharie Jehl, Edgardo Saucedo, Claudio Cazorla