Quantum Physics
Classical Bound on the Fisher Information Rate of a Dephasing-Enhanced Quantum Neuron
Peng Wang, Yu-Xuan Zhang, Hai-Tao Ding et al. (+1 more)
Dephasing can sharpen the activation of a dissipative
quantum neuron, but a stationary response does not determine how much input information its output delivers per unit time. For a single
qubit with partial-reset updates at rate $\nu$, we derive the threshold response bandwidth and the exact local Fisher information rate of the complete record from repeated projective readouts, including their backaction. Near the activation threshold, the rate obeys $\mathcal R\leq\nu(1-s)(1+s)^2/2\leq16\nu/27$, where $s$ is the retained amplitude of each update. For every fixed finite Markovian dephasing rate, optimization over the readout interval and $s$ gives the same supremum, approached at $s=1/3$ with arbitrarily rapid ideal readout. A finite measurement dead time instead selects a finite optimal interval, which we obtain by a dimensionless two-parameter optimization. The ideal bound is saturated by a classical two-state jump process:
quantum coherence changes finite-cadence performance but cannot raise the optimized bound. These results separate noise-enhanced activation from the operational information throughput of a
quantum neuron and provide a hardware-aware
benchmark for finite-time
quantum neural processing.
Read the full abstract ▾
All authors Peng Wang, Yu-Xuan Zhang, Hai-Tao Ding, Leong-Chuan Kwek
Machine Learning
ReCo: When to Relocate Sensor Kits under Deployment Constraints – A NILM Case Study
Haokun Chen, Yu Tong and Yehai Chen
Many sensing tasks obtain training labels only by deploying instruments in the field. With a limited number of sensor kits, a collection deadline, and measurement downtime at every move, the collector must repeatedly decide whether to stay at the current site or relocate. We study this decision in non-intrusive load monitoring (NILM), which estimates the power drawn by individual appliances from a home's main meter and is trained on data from homes temporarily fitted with appliance-level sub-meters. In NILM, appliance usage varies with the appliance, season and
climate, and the value of new data depends on how diverse the combinations of target operation and background load are. To address this, we propose a constraint-based relocation framework and instantiate it for NILM as ReCo (Relocation by Coverage gain). ReCo counts new operating regimes in a joint target-background feature space, forecasts each home's future gain from the data collected so far, and each night weighs the gain of staying against the gain of moving elsewhere after the downtime. In replayed deployments on the Plegma
dataset under two kit counts and two downtime costs, ReCo outperforms fixed-dwell and count-based schedules and a threshold rule using the same metric in every setting. Its advantage is not explained by collecting more days alone and reflects allocating the days to more valuable homes and periods.
Read the full abstract ▾
Computation and Language
When Confidence Rises Too Early: Detecting Shortcut Reasoning via Premature Answer Commitment
Zhaohan Zhang, Junjie Liu, Chengzhengxu Li et al. (+6 more)
The
reasoning trajectory of a Large
Language Model (LLM) is often treated as a verbalized description of its internal
reasoning. However, such trajectories can be unfaithful: a model may rely on shortcuts to reach an answer and then post-rationalize the decision with a seemingly coherent chain of thought. Detecting this shortcut
reasoning is challenging because existing monitors and verifiers mainly inspect textual traces or final outcomes, rather than how the model's belief in its answer develops during generation. We introduce ConfLens, a framework that tracks the
evolution of confidence in the final answer throughout
reasoning. Across three shortcut
reasoning settings, we observe a common pattern of premature confidence, where shortcut samples become highly confident in the final answer at early
reasoning stages. Existing confidence estimation methods, however, show limited generalizability, reliability, or efficiency for detecting this behavior. We therefore propose the Distributional Answer Commitment Score (DACS), a distributional confidence
estimator that measures the
entropy of the model's probability distribution over answer commitment at each
reasoning step. DACS captures how concentrated the model's answer belief is without requiring ground-truth answers or task-specific verifiers. We further convert ConfLens detection results into interpretable signals for reward models to reduce their preference for shortcut
reasoning. Experiments on mathematical and code
reasoning tasks show that ConfLens with DACS improves shortcut
reasoning detection by over 4.3% F1 compared with strong baselines and reduces the mismatch between faithfulness and correctness in reward model preferences.
Read the full abstract ▾
All authors Zhaohan Zhang, Junjie Liu, Chengzhengxu Li, Chen Shen, Xiaoming Liu, Chao Shen, Jieping Ye, Ziquan Liu, Ioannis Patras
Machine Learning
Interrelating Fruchterman-Reingold Graph Visualization and Agglomerative Clustering
Alexandre Benatti and Luciano da F. Costa
Graph visualization methods and agglomerative clustering have been frequently considered in data analysis and pattern recognition. Because these approaches are interrelated and complementary, it is of particular interest to investigate their associations. In this work, we study the possible relationship between the Fruchterman-Reingold graph visualization method and four types of agglomerative clustering adopting single- and complete-linkage, average, and Ward's linkage criteria. Three types of
datasets have been considered in 2 and 10 dimensions, as well as the PCA projection of the latter to two dimensions. The results obtained suggest that the relationship between the methods considered did not vary much for the three types of data mentioned above. At the same time, the agglomerative methods tended to yield results that are mostly similar to each other, while presenting moderate similarity with the original data. The Fruchterman-Reingold visualization resulted similar to the original data, but exhibited relatively smaller similarity to the agglomerative methods.
Read the full abstract ▾
Machine Learning
Not All Rollouts Are Worth Learning: On Trajectory Valuation for Post-Training Reinforcement Learning
Xuesong Jia, Ziao Yang, Zhanhe Huang et al. (+1 more)
We consider the problem of trajectory valuation in reinforcement learning: how to identify and mitigate detrimental trajectories during online training. Unlike classification, where data valuation relies on fixed training and validation sets, reinforcement learning involves dynamically generated trajectories without explicit validation signals, making conventional influence-based methods inapplicable. We propose Dynamic Trajectory Valuation (DTV), a simple and efficient framework that estimates trajectory utility at the mini-batch level and filters detrimental trajectories based solely on gradient information. By operating at the optimization level, DTV integrates seamlessly with existing reinforcement learning pipelines with minimal overhead. Extensive experiments across diverse settings, including PPO, GRPO, and DPO, demonstrate that DTV consistently improves performance, enhances data efficiency, and stabilizes optimization.
Read the full abstract ▾
All authors Xuesong Jia, Ziao Yang, Zhanhe Huang, Hongfu Liu
Differential Geometry
Derived deformation theory of heterotic $G_2$ systems near the standard embedding
Bram Brongers
We study heterotic $G_2$ structures near a torsion-free standard
embedding, to first order in $\alpha'$ and to all orders in formal deformations. The tangent connection is the Hull connection determined by the metric and flux. We construct a locally invertible family of differential operators relating the supersymmetry and instanton $BPS$ equations to the Euler-Lagrange equations of the heterotic superpotential. Using this comparison and the physical gauge
symmetries, we construct a homological vector field locally formally conjugate to the
Hamiltonian vector field of the superpotential. On a closed
manifold in the fixed gerbe sector, the torsion equation implies vanishing of the leading flux. We prove ellipticity of the linearized deformation complex with mixed differential-order weights and construct a finite-dimensional cyclic $L_\infty$ model with a formal effective potential.
Read the full abstract ▾
Computation and Language
Echoes of Deeds: Moral History Can Shape and Steer LLM Behavioral Choices
Lucio La Cava and Andrea Tagarelli
Evaluations of Large
Language Models (LLMs) morality typically consider decisions in isolation, thus overlooking whether an individual's unrelated prior conduct influences the model's subsequent choices. This leaves open the question of whether, and to what extent, moral history shapes LLM decisional behaviors. Prior work on human moral decision-making shows that past behavior can influence subsequent moral choices. Building on this observation, we investigate whether analogous effects emerge in LLMs in two complementary ways: at the behavioral level, through the model's observable responses, and at the representation level, through its latent internal representations. We introduce MoralLedger, a framework for studying how an actor's moral history shapes actions for LLMs' behaviors under a fixed decision context. At the behavioral level, we find that prior moral histories systematically alter subsequent choices as a function of their valence and intensity. At the internal representation level, these histories induce a linearly recoverable direction in the residual stream that generalizes to held-out examples. Intervening along this direction on neutral-history
prompts produces two-sided intensity-dependent changes in subsequent choices, with effects that are stronger than those induced by prompting alone or by favorable-nonmoral direction. To our knowledge, this is the first demonstration that a latent representation of an actor's prior moral conduct can provide signed inference-time control over a moral decision. Our MoralLedger extends moral evaluation beyond static dilemmas, establishing moral history as both a source of behavioral sensitivity and a causal target for
auditing and controlling moral behavior in LLMs.
Read the full abstract ▾
Information Retrieval
RenderRank: Learning to Rerank Text with Compressed Visual Tokens
Seongtae Hong, Youngjoon Jang, Jungseob Lee et al. (+2 more)
Rendering document text as images allows
vision-
language models to encode documents as visual
tokens, which can reduce input sequence length compared with text input. This reduction in input length is particularly useful for reranking, where each query involves scoring multiple candidate documents and
token savings apply to each candidate evaluation. We introduce RenderRank, a reranker that learns query-dependent relevance scoring from compressed visual document representations instead of the text
token sequences used by conventional text-based rerankers. Training first aligns relevance scores from visual inputs with those of a text-based teacher, then refines the relative scores of positive and negative documents for the same query. Across 11
datasets from BEIR, RenderRank uses 16.5-35.5% fewer input
tokens while achieving an average NDCG@10 of 55.96, outperforming all evaluated text-based baselines below 4B parameters and some larger models. Across four long-document
datasets, it achieves an average NDCG@10 of 88.27 with approximately half the average input
token count of the evaluated text-based rerankers. In this setting, RenderRank delivers 1.70x the highest average throughput of the evaluated baselines. These results demonstrate that compressed visual representations can support accurate document relevance scoring, providing an alternative to text
token representations for reranking.
Read the full abstract ▾
All authors Seongtae Hong, Youngjoon Jang, Jungseob Lee, Hyeonseok Moon, Heuiseok Lim
Soft Condensed Matter
Electrolyte density, diffusivity and conductivity in graphene nanoconfinement predicted by separating interfacial from genuine confinement effects
Haoyuan Quan, Hanne S. Antila, Maximilian R. Becker et al. (+2 more)
Confined aqueous electrolytes exhibit strong deviations from bulk behaviour, but it remains unclear which changes arise from genuine confinement-induced modification and which merely reflect the influence of interfaces. Here, we introduce an interfacial deficit-length framework, capable of decomposing density, diffusivity, and conductivity of confined electrolyte solutions into interfacial and confinement contributions. By applying the framework to molecular dynamics simulations of aqueous alkali halides in planar
graphene nanoslits coupled to bulk reservoirs at variable electrolyte concentration, we show that genuine confinement effects emerge only for slit heights $H \lesssim 1$ nm; for larger H, deviations from bulk behaviour are quantitatively captured by interfacial deficit lengths. These deficit lengths are strongly ion-specific and generally positive for water and salt densities as well as for conductivities, meaning that
graphene interfaces reduce the values of these observables, while diffusivity deficit lengths tend to be negative, corresponding to larger slit self-diffusivities relative to the corresponding bulk reference. Our deficit-length framework is applicable to any observable from experiments or simulations on nanoconfined electrolytes and predicts confined electrolyte properties for variable slit height H and reservoir electrolyte concentrations.
Read the full abstract ▾
All authors Haoyuan Quan, Hanne S. Antila, Maximilian R. Becker, Philip R. Loche, Roland R. Netz
Quantum Algebra
Isomorphism between the R-matrix and Drinfeld presentations of quantum affine superalgebra for type $\boldsymbol{A}$
Pengfa Xu, Hongda Lin and Honglian Zhang
In this paper, we establish an explicit isomorphism between Drinfeld and R-
matrix presentations of the
quantum affine superalgebra associated with the general linear Lie superalgebra $\mathfrak{gl}(m|n)$. This result can be viewed as a supersymmetric analogue of the isomorphism between two presentations of the
quantum affine
algebra $U_{q}(\widehat{\mathfrak{gl}(n)})$ proposed by Frenkel-Ding. Furthermore, employing the
quantum Berezinian of $U_{q}(\widehat{\mathfrak{gl}(m|n))}$, we construct the R-
matrix presentation of the
quantum affine superalgebra $U_{q}(\widehat{\mathfrak{sl}(m|n))}$ and prove the isomorphism between its R-
matrix and Drinfeld presentations. Our proof based on the commutative relation between Gaussian generators, $\mathbb{A}$-form and
quantum Berezinian.
Read the full abstract ▾
Machine Learning
Depot-Closed Multi-Component Construction for Neural Vehicle Routing
Shinichiro Hamada and Hisashi Kashima
Most neural constructive solvers for the vehicle routing problem (VRP) use route-by-route construction, extending one route until completion before starting the next. This commits route membership early and hinders global coordination across routes. We propose multi-component construction, which maintains many route components simultaneously and merges them in an arbitrary order. This removes the depot-return cue that route-by-route construction obtains from the remaining capacity; to compensate, we introduce an interpretation in which every component is treated as an implicitly depot-closed route. Under this depot-closed interpretation, every intermediate state of standard CVRP construction is a complete feasible solution, and the exact cost reduction of a merge is the Clarke-Wright saving. The neural policy combines this CW-saving signal with the evolving component state to learn what to connect and when to connect. A policy trained only on CVRP100 outperforms the reported results of representative neural solvers on CVRP100-500 with greedy inference and, reused for ruin-and-reconstruct, performs strongly at all evaluated sizes up to CVRP1000. In a zero-shot Constraint Tightness evaluation with capacities from $C=10$ to $500$, it outperforms the reported neural solvers at every capacity. Controlled analyses show that robustness persists without CW grounding and point to learned route-closing behavior as a plausible contributor to the tight-regime degradation of learned route-by-route solvers.
Read the full abstract ▾
Distributed Computing
TempoKV: Timely Staging of LLM KV Caches for Memory-Semantic Flash
Jay H. Park, Hyungjun Kim and Dong Kim
Reusable prefix key-value (KV) caches can outgrow GPU
memory in large
language model (LLM) serving. A
memory-semantic flash hierarchy offers SSD-backed capacity with a limited fast tier, but a logical KV hit is not necessarily ready for GPU
retrieval. Demand staging exposes SSD latency, whereas immediate staging can reserve fast-tier capacity long before
retrieval begins. We present TempoKV, a timing-aware resource-commitment layer that separates early knowledge of reuse from the acquisition of staging resources. It records reusable-KV hits as metadata-only claims and requests commitment when the runtime-estimated time until
retrieval falls to the storage-estimated time needed to make KV resident and protected against eviction. These estimates adapt to runtime progress and staging state, while commitment remains subject to available protected capacity. We implement TempoKV in vLLM and LMCache on an SSD-backed CXL
memory device without changing request scheduling. Across two models and three prefix cache ratios, TempoKV reduces protected fast-tier byte-time per request by 63-91% versus immediate staging while retaining much of the serving benefit of advance staging. In a fast-tier capacity sweep, output throughput and p95 time to first
token (TTFT) remain nearly unchanged as capacity decreases from 100 to 25 GiB. Compared with unmodified LMCache's Device-DAX L1 configuration, TempoKV reduces p95 TTFT by up to 48.0% and increases output throughput by up to 27.8%.
Read the full abstract ▾
Instrumentation and Methods for Astrophysics
I2C@2024: an interferometric imaging contest in 2024
Florentin Millour, Mathis Houllé, Jeremy Perdigon et al. (+6 more)
Images in optical long-baseline interferometry have seen a boost in the recent years thanks to new techniques and recipes invented by the community. These images are more and more used for science interpretation and not only illustration, and their fidelity has improved significantly, thanks mainly to the increase in the number of
telescopes used in interferometers. The focus today is to improve their reliability and dynamic range. With this contest, we follow up the quest introduced in 2004 of comparing the state of the art image reconstruction software for long-baseline interferometry. This is done in a festive way in the form of an imaging contest, where the organizers propose simulated
datasets of targets, whose brightness distributions are meant to be blindly retrieved using various
algorithms by the contestants. A prize is offered to the winner of the contest. This year is not different from previous ones and we proposed to the contestants tools to compare their reconstructed images with original images. These tools are now distributed, together with example
datasets and images, enabling further tests at home of any image reconstruction tool.
Read the full abstract ▾
All authors Florentin Millour, Mathis Houllé, Jeremy Perdigon, Julien Drevon, Ryan Norris, Rebecca Proni, Alexis Matter, Ferreol Soulez, Eric Thiebaut
Mathematical Finance
Optimal Liquidation with Support and Resistance Levels under Multi-Skew Brownian Motion
Jun Maeda
We solve the perpetual liquidation problem for a
geometric multi-skew Brownian motion carrying local-time pushes upward at a support level and downward at a resistance level, a model of technical analysis that is Markov in the price alone. Three geometries arise, separated by a closed-form criterion: the continuation band lies below resistance, straddles it, or pins its upper edge there. Selling into resistance is a conclusion rather than an assumption. Identification is asymmetric: exercise behaviour determines the support parameters exactly, while the strength of resistance is recoverable only on an interval with a closed-form endpoint.
Read the full abstract ▾
Theoretical Economics
Blackwell Boundaries
Shuo Li Liu
We give a finite simplex
theorem. When the number of states equals the number of source signals and the source posterior likelihood-ratio vectors form a simplex, Blackwell dominance is characterized by a universal barycentric-coordinate condition. We also give an explicit countable-state, countable-signal diagnostic-monitor
theorem with a closed-form universal criterion and a closed-form garbling; tensorization yields dominance at every sample size. Finally, we provide a three-state, two-signal counterexample to the sufficiency conjectured by Mu, Pomatto, Strack, and Tamuz for their many-state moment-generating-function conditions for large-sample Blackwell dominance. The pair satisfies strict comparisons on both parameter domains, all ordered Kullback-Leibler inequalities, bounded likelihood ratios and pairwise genericity, but dominance fails at every positive sample size. The finite
theorem, obstruction, and continuum inequalities are verified in Lean 4.30 with Mathlib; the countable diagnostic proof is given analytically.
Read the full abstract ▾
Information Theory
Improved bounds for constant-power and low-power error-correcting cooling codes
Tingting Tong and Sihuang Hu
Low-power error-correcting cooling (LPECC) codes and constant-power error-correcting cooling (CPECC) codes provide error correction while controlling power consumption and thermal effects in on-chip buses. In this paper, we study binary CPECC and LPECC codes with \(e=w-3\). For CPECC codes, we extend the applicability of the upper bound previously obtained by Zhao and Zhang from the quadratic-order condition \(w\ge 2t(t+1)+2\) to \(w\ge w_0(t)\), where \(w_0(t)\sim \sqrt{2}\,t^{3/2}\). Using Steiner systems, we show that the CPECC bound is attainable and asymptotically tight for fixed \(t,w\). For LPECC codes, we establish the new upper bound \(\left\lfloor\frac{\binom{n+2}{3}}{\binom{w+t}{3}}\right\rfloor\) for \(w\ge \mu(t)\), where \(\mu(t)\sim \sqrt{2}\,t^{3/2}\). This bound is strictly smaller than the previous bound of Zhao and Zhang whenever both apply, and is asymptotically tight for fixed \(t,w\) in the stated range.
Read the full abstract ▾
Machine Learning
Beyond Gradient Flow: Identifiability and Recovery from Distribution Snapshots
Nam D. Nguyen and Valeriya Malysheva
Inferring dynamics from snapshots of evolving distributions is fundamentally underdetermined: the Fokker-Planck equation constrains the drift $F$ only through its score-weighted divergence $\nabla\cdot F+F\cdot\nabla\log\rho$, leaving a $\rho$-solenoidal gauge invisible to any single-time constraint. Time-indexed transport formulations cannot resolve this ambiguity: every admissible marginal path admits a curl-free explanation, minimum-action reconstruction selects it, and marginal fit alone cannot distinguish dynamically inequivalent explanations. Requiring one autonomous field to explain several marginals instead makes part of the hidden circulation visible as $\nabla\log\rho$ changes across marginals. Separating instantaneous Fokker-Planck source constraints from the snapshot experiment, we show that the source constraints identify the field modulo the kernel of a stacked score-weighted divergence operator. For generic Gaussian shape variation, source constraints at $K\ge m$ time points in intrinsic dimension $m$ eliminate every
polynomial gauge direction, whereas finitely many density snapshots alone admit aliasing; we give the obstruction explicitly. At a Gaussian anchor, for Sobolev smoothness $s$ and $n$ samples per time point, we derive a conditional lower rate $(nK)^{-2s/(2s+m+1)}$ for the tangent snapshot experiment, with a matching upper rate in a degreewise
benchmark. Strong-form fitting is non-orthogonal to score error and cannot be repaired by spectral filtering. Instead, we estimate using smooth test functions while retaining the known
diffusion term, and derive a finite-sample bound that separates sampling error from fixed-grid quadrature bias. Planted-circulation experiments confirm the predicted gauge contraction and expose a design tension between cross-slice information and covariance-aware whitening.
Read the full abstract ▾
Computer Vision
Towards Generalizable 3D Anomaly Detection via Relational Inconsistency Modeling
KunHo Heo, SuYeon Kim, Hayoung Lee et al. (+2 more)
3D anomaly detection (3DAD) aims to identify defective regions in point cloud data, serving as a critical component in industrial inspection systems. Existing methods are normality-centered – learning the distribution of normal samples and treating deviations as anomalies – without explicitly modeling what constitutes a defect. This leads to ambiguous decision boundaries with increased false positives and negatives, particularly in unified and cross-domain settings where diverse normal distributions further blur the boundaries. We propose a relational inconsistency modeling framework that characterizes defects as violations of
geometric consistency among neighboring structures. Our approach learns category-agnostic defect cues through pseudo-anomalies designed as controlled relational violations, instantiated by two key modules: Edge-aware Graph Refinement (EGR) for encoding
geometric relationships among local regions, and Cluster-Deviation Modeling (CDM) for identifying regions that are relationally incompatible within their structural peer group. Extensive experiments on Anomaly-ShapeNet and Real3D-AD demonstrate consistent improvements over prior state-of-the-art methods in both in-domain and cross-domain settings, validating the effectiveness of learning an explicit, relation-based defect criterion for 3D anomaly detection. Project page: https://visualsciencelab-khu.github.io/GRIM_project/.
Read the full abstract ▾
All authors KunHo Heo, SuYeon Kim, Hayoung Lee, Chanse Oh, MyeongAh Cho
Machine Learning
TIDE: Teacher-Student Transition via Informative Distillation and Exploration for Agentic RL
Yibin Huang, Xinming Xu and Conghui Zhu
Effective multi-turn
agents require interaction strategies that coordinate information gathering, actions, and feedback over long horizons. GRPO is a reinforcement learning
algorithm used to train these
agents, but sparse trajectory-level rewards limit early exploration in small models. Recent methods augment RL with on-policy
distillation (OPD) from a stronger teacher. However, a fixed mixture assumes that teacher guidance and reward optimization should retain a constant relative role throughout training and across interaction turns. This assumption can fail at two scales. Globally, as training progresses, maintaining strong
distillation pressure can constrain the model from moving beyond the teacher's capabilities. Locally, teacher–student disagreement identifies where the student departs from the teacher, but cannot tell whether that departure is exploration supported by better outcomes or low-quality policy drift. Our methodological insight is that teacher guidance and reward optimization should be dynamically rebalanced over training and jointly allocated across turns. We instantiate this insight in \tide. Globally, \tide uses the measured disagreement trend as a practical schedule signal, advancing an OPD-to-RL handoff when discrepancy reduction becomes slow but remains positive and progressively increasing the relative weight of RL. Locally, \tide jointly modulates teacher-guided and reward-driven updates: relative action value and disagreement prioritize the OPD signal, whereas relative action value supplies the RL advantage and normalized disagreement reweights it across turns. Coupled with the global handoff, \tide allocates stronger teacher guidance early and gives reward-driven updates greater relative weight later in training. Experiments across multiple
benchmarks, student scales, and controlled ablations support the effectiveness of TIDE's adaptive OPD–RL coordination.
Read the full abstract ▾
Cosmology and Nongalactic Astrophysics
MATISSE in the era of GRAVITY+
Florentin Millour, Anthony Soulain, Stéphane Lagarde et al. (+28 more)
MATISSE, the mid-infrared spectro-interferometric imager of the Very Large
Telescope Interferometer (VLTI), has been feeding on the light corrected by the
GRAVITY+ Adaptive Optics (GPAO) systems since the retirement of MACAO on the four Unit
Telescopes in 2024. We report the assessment of the new MATISSE performance with GPAO, based on commissioning observations obtained in 2024 and 2025 and compared with archival MACAO observations of the same
stellar sources. The instrumental visibility transfer function is more stable in the L band, while its typical value in the L and N bands is unchanged; the visibility signal-to-noise ratio is consistent with, or better than, the values reported at the initial MATISSE commissioning. The correlated flux increases by 20–30% in GRA4MAT mode thanks to a better fringe tracking performance, and the total flux injected into the instrument increases by a factor of 1.3 to 1.4 (up to 1.7) in the L band at low spectral resolution, with no measurable loss in the N band. Combined with the new K-band fringe-tracking limits enabled by GPAO, this translates into a sensitivity gain of a factor 2 to 6 in the L and M bands, a confirmation of the N-band limits similar to the MACAO era. We then report the performances of MATISSE with the GPAO's
laser guide
stars in November and December 2025. The updated noise figures follow the exposure time calculator predictions in LM LOW and MED and come out better than them in N LOW, the L-band transfer function is stable at the few percent level between observing blocks an hour apart, and correlated fluxes, differential phases and closure phases are recovered down to a 10\,mJy target. Embedded young
stellar objects and AGN are now within reach, thanks to the better performances on red targets brought by the
lasers. We finally discuss the science cases opened by this gain and the case for a future upgrade of the instrument.
Read the full abstract ▾
All authors Florentin Millour, Anthony Soulain, Stéphane Lagarde, Alexis Matter, Valentin Fleury, Farin Drewes, Mathis Houllé, Jules Scigliuto, Bruno Lopez, Anthony Meilland, G. Petrov Romain, Philippe Berio, Anthony Berdeu, Guillaume Bourdarot, Sylvestre Lacour, Mathias Nowak, Jean-Baptiste Le Bouquin, Antoine Mérand, Frank Eisenhauer, Julien Woillez, Miguel Montargès, Diogo Ribeiro, Taro Shimizu, Jens Kammerer, Xavier Haubois, Stefan Gillessen, Pierre Cruzalèbes, Frédéric Gonté, Walter Jaffe, James Leftley, Nicole Nesvadba
Neurons and Cognition
Scaling Laws for EEG Decoding: How Much Data Is Enough?
José Maurício Nunes de Oliveira, Bruna J. Lopes, Léo Burgund et al. (+2 more)
Deep learning has become a cornerstone of
EEG-based
brain decoding, with a growing number of architectures proposed every day. However, how the performance of these different models scales with data volume is not clear. Although this relationship has been characterized in other fields under the name of scaling laws, it remains poorly understood in the
EEG domain. The present study addresses this gap by investigating how scan time and subject diversity affect the performance of different architectures. We evaluated five models across four
EEG datasets. Training data volume was controlled by varying both subject count and trial volume under cross-subject validation. We then fitted power-law relationships to characterize the resulting behavior. Our findings reveal that as total data volume increases, the distinction between trial and subject scaling becomes largely irrelevant. Furthermore, we show that power-law relationships are both model and
dataset-specific, yet they provide a robust descriptive framework for
EEG decoding performance. Extrapolation to larger subject pools yields RMSE values below 0.1 in most cases. Our work contributes to the literature by providing a descriptive framework for data scaling in
EEG and by demonstrating data-efficient experimental design in
EEG research.
Read the full abstract ▾
All authors José Maurício Nunes de Oliveira, Bruna J. Lopes, Léo Burgund, Raphael Y. Camargo, Bruno Aristimunha
Machine Learning
Universality and Generalization of Causal Transformers Across Context Lengths
Takashi Furuya, Maarten V. de Hoop and Gabriel Peyré
Long contexts are central to modern
transformer systems, but most expressivity results choose a different network for each fixed sequence length. We study whether one masked
transformer can approximate causal
token-to-
token maps uniformly over sequences of arbitrary length sampling a fixed normalized horizon. To relate sampling resolutions, we model
tokens by $\alpha$-Hölder sequences or, more generally, a common modulus of continuity. Our notion of continuity across resolutions characterizes the causal families admitting uniform approximation on these compact input classes by a single
transformer with length-independent parameters. The result extends to the infinite-length mean-field limit, where
tokens form continuous curves and masked
attention becomes a causal time integral. For bounded regression with target maps satisfying a $\beta$-smooth stability condition defined using regular test functions, quantitative approximation yields a generalization bound: exact empirical risk minimization over suitably sized bounded-weight
transformers gives root mean-square prediction error $O((\log\log N/\log N)^{\beta/(d+2)})$ from $N$ iid labeled sequences. The bound holds at fixed confidence on the same sampling distribution, with $d$ the
token dimension and no maximum-length factor. Finally, experiments on physical time series support the Hölder-regular
token model at observed scales, with
dataset-dependent fitted exponents, whereas text input
embeddings provide a contrasting case. Native and dense sampling, shuffled controls, and refinement checks delimit this empirical regularity regime.
Read the full abstract ▾
Audio and Speech Processing
Perceptual Quality Loss or Loss of Perceptual Quality?
Danilo de Oliveira, Tal Peer, Maurício do V. M. da Costa et al. (+1 more)
Contemporary deep
speech enhancement (SE) models are often trained with specific auxiliary terms in the loss function as a way to improve their performance in terms of perceptual metrics. Nevertheless, a higher score on a perceptual metric does not necessarily correlate with an improved listening experience. Through objective and subjective experiments, we assess the performance of SE models trained with two different types of auxiliary PESQ loss terms. The numerical evaluation on a suite of standard metrics suggests that, while models optimized for PESQ naturally obtain higher PESQ scores in the test set, for most other metrics the scores do not significantly change. In some cases, the PESQ loss even results in worse PESQ scores on mismatched data. A formal listening experiment reveals that the models without a PESQ loss were generally preferred over models that include it, across all settings. Finally, we analyze the relative importance of PESQ in the composite metrics CSIG, CBAK and COVL, and find that PESQ dominates all of them. Our study highlights the perils of over-reliance on PESQ and stresses the importance of a complete evaluation procedure for SE.
Read the full abstract ▾
All authors Danilo de Oliveira, Tal Peer, Maurício do V. M. da Costa, Timo Gerkmann
Cryptography and Security
Efficient TCitH-Based Alternatives to SLH-DSA: Cross-Layer ASIC Design of Mirath
Hiandra Tomasi, Maximilian Schöffel, Johannes Feldmann et al. (+1 more)
To address the security risks posed by
quantum computers, the U.S. National Institute of Standards and Technology (NIST) has standardized the post-
quantum signature schemes ML-DSA, FN-DSA, and SLH-DSA. While ML-DSA and FN-DSA are
lattice-based, SLH-DSA relies on hash-based assumptions. To support cryptographic agility against future vulnerabilities, NIST is evaluating non-
lattice candidates as alternatives to SLH-DSA. Among these, TCitH-based schemes are particularly promising due to their compact keys and small signatures. However, their high computational complexity and
memory footprint pose significant challenges for efficient implementations on resource-constrained embedded platforms. They remain largely unexplored in this context, particularly in ASIC implementations. To address this gap, we use a cross-layer methodology combining algorithmic and hardware layers to present, to the best of our knowledge, the first ASIC implementation of Mirath, a TCitH-based signature scheme, in a RISC-V-based system. The design is implemented in a 22nm FD-SOI technology node. Compared with an SLH-DSA ASIC implemented in the same technology node, the proposed architecture achieves 17.8x lower signing latency while requiring 58% less total
cell area, showing the potential of TCitH-based signatures as efficient non-
lattice alternatives from an implementation perspective.
Read the full abstract ▾
All authors Hiandra Tomasi, Maximilian Schöffel, Johannes Feldmann, Norbert Wehn
Computer Vision
OPIS: An Input-Grounded Benchmark for Multi-Object Memory in Video World Models
Hao Wang, Tao Yu, Liuzhou Zhang et al. (+13 more)
Video world models must preserve the visual state of the world over time, but existing evaluation protocols often rely on generated histories,
video reference, or selected revisit viewpoints that can confound the assessment of a model's true
memory capability. To address this, we introduce OPIS, an input-grounded
benchmark that strictly anchors the assessment to a fixed set of object instances from the initial observation for evaluating multi-object
memory in
video world models. The OPIS
dataset comprises 500 cases across real-world,
embodied-robotic, and game-world domains, providing dense object-level annotations for 12,672 rigid, articulated, and deformable instances. Our object-centric evaluator combines association and explicit visibility
reasoning to hierarchically measure Object (O) Presence (P), Identity (I), and Structure (S), utilizing static or dynamic evaluation tracks based on object kinematics. Across eight image-to-
video or camera-conditioned world models, our proposed OPIS scores range from 48.65 to 56.01. As the reference inventory grows from less than 20 to more than 40 objects, the Presence, Identity, and Structure scores show an overall decline, with the average Identity score falling from 40.22 to 23.11. The results demonstrate that preserving the particular object instances in the input is considerably harder than generating plausible visual elements.
Read the full abstract ▾
All authors Hao Wang, Tao Yu, Liuzhou Zhang, HeXin Wang, Haopeng Jin, Yuxuan Zhou, Xinming Wang, Hongzhu Yi, Xinye Li, Yuanlei Wang, Ping Nie, Yan Huang, Yuxuan Zhang, Pengfei Zhou, Yanyan Zou, Wei Yang
Information Theory
Large-Scale Autonomous Discovery of Kissing Number Constructions
Shuxing Yang, Rui Zhao, Junyao Wu et al. (+13 more)
The kissing-number problem is a classical problem in discrete geometry whose exact solution is known in only a few dimensions. Recent artificial-intelligence approaches have begun to discover improved configurations through large-scale numerical and combinatorial search, but converting such searches into general mathematical constructions and rigorous proofs remains challenging. Here we use Qiushi Engine, an autonomous multi-
agent research system, to investigate kissing numbers and obtain new lower bounds in nineteen dimensions: $25$, $27$, $32$–$39$, $43$, $45$, and $49$–$55$. The resulting constructions arise from distinct structural mechanisms, including coordinated motions of contact layers, labelled direction reuse, joint support exchanges, signed-code replacements, cross-shell
lattice constructions, low-overlap
lattice isometries, and spherical-design moment certificates. They yield sharp capacities for parameterized signed-code models, deterministic image-union guarantees, and exact section and projection counts controlled by embedded root systems and anchor-graph statistics. These methods yield, among others, $K(25)\ge197580$, $K(27)\ge201567$, $K(38)\ge591900$, $K(43)\ge2553792$, $K(45)\ge7380090$, and $K(55)\ge53301140$. The autonomous system carried out the construction searches, mathematical analysis and computational verification, while each final result was reduced to explicit mathematical arguments and independently checkable finite certificates. Our results illustrate how autonomous research systems can move beyond optimization within fixed formulations to discover new mathematical representations and constructions at scale.
Read the full abstract ▾
All authors Shuxing Yang, Rui Zhao, Junyao Wu, Yize Wang, Fujia Chen, Kaihao Zhu, Wenhao Li, Zichen Li, Yaqi Li, Shenzhan Hong, Yuang Pan, Junjie Yang, Taowen Deng, Jincheng Mi, Hongsheng Chen, Yihao Yang
Fluid Dynamics
Handling geometric singularities efficiently with the immersed-boundary method: application to riblets
Stefano Cipelli, Alessandro Chiarini, Maurizio Quadrio et al. (+2 more)
We present an enhancement of classical immersed boundary methods for the direct numerical simulation of the incompressible Navier-Stokes equations, designed to efficiently cope with geometrical singularities. The method, while general, is described here in the context of the flow over a plane wall covered by small longitudinal riblets. The enhancement consists in prescribing that the solution near the boundary matches the steady Stokes solution. In this way, sharp corners, where velocity gradients are strongest and viscous effects dominate, become tractable with grids comparable to those necessary for a smooth wall. The method requires the availability of a local Stokes solution, that can be obtained either analytically or numerically. Two test cases are discussed: the laminar flow over one riblet, which allows to measure protrusion heights, and a fully turbulent channel flow with ribbed walls. In both cases, the method reproduces reference results at a substantially lower computational cost.
Read the full abstract ▾
All authors Stefano Cipelli, Alessandro Chiarini, Maurizio Quadrio, Davide Gatti, Paolo Luchini
Instrumentation and Methods for Astrophysics
The best of all worlds: gravitational-wave transient detection with multiple pipelines
Nikolas Moustakidis, Theofilos Moustakidis, Deep Chatterjee et al. (+2 more)
The search for gravitational-wave transients (modeled and unmodeled) as performed by the LIGO-Virgo-KAGRA collaborations and the broader community typically involves multiple detection
algorithms, often redundant and sometimes complementary. We here address the problem of combining such detection
algorithms in order to maximize the noise vs signal discrimination power of such an astrophysical search that utilizes multiple analysis pipelines. Using a machine learning approach, we construct a neural network-based classifier that utilizes information (or lack thereof) from all participating pipelines in the search for compact binary coalescences in LIGO-Virgo-KAGRA's realtime (and offline) processing of gravitational-wave data. We
benchmark the method with simulated (astrophysical) signals and real noise from the instruments and obtain receiver operating characteristic curves to quantify its performance. We show how this leads to an effectively new, combined pipeline that outperforms any single pipeline alone or other logical combinations of them. This is particularly critical for real-time operations of the LIGO-Virgo-KAGRA detectors as such combination of detection methods can lead to increased robustness in identifying astrophysical sources and to reduction of the number of false alerts that may otherwise be sent out as astronomical telegrams. Our method also aims in simplifying the presentation of astrophysical results out of the LIGO-Virgo-KAGRA detectors by combining all participating pipelines in the astrophysical searches as well as optimizing resources by the broader multi-messenger astrophysics community in following up astronomical alerts for gravitational-wave transient events.
Read the full abstract ▾
All authors Nikolas Moustakidis, Theofilos Moustakidis, Deep Chatterjee, Anastasios Tefas, Erik Katsavounidis
Symplectic Geometry
Non-decomposable cobordisms via Lagrangian moves
Roberta Guadagni
We establish the use of Lagrangian diagram moves to build Lagrangian cobordisms of Legendrian
knots. We then apply the moves to explore how to invert decomposable cobordisms, after adding some stabilizations. This is done by use of Reidemeister moves and handle attachments in the Lagrangian projection. As an application, we show an explicit construction of a Lagrangian concordance from a stabilized $m(9_{46})$ Legendrian
knot to a stabilized unknot, i.e. a non-decomposable Lagrangian concordance of nonempty
knots.
Read the full abstract ▾
Robotics
EMPIRIC: Experiment-Driven Learning of Residual World Models for Robot Planning
Yichao Liang, Amber Li, Dat Nguyen et al. (+13 more)
A
robot should be able to learn through experiments how unfamiliar objects behave and interact, then plan with that knowledge. It need not start from scratch: physics engines supply knowledge of motion and contact, but can omit entire mechanisms, such as glue curing, water heating, or wind. We present EMPIRIC, an
agent that learns a residual world model: a physics engine extended with code for the missing mechanisms. The learned programs can introduce new forces, constraints, and hidden state, and
Bayesian inference estimates their parameters and states from noisy observations. The resulting model lets the
agent predict the outcomes of actions, choose informative experiments, and revise its hypotheses when predictions fail. Across five simulated domains, EMPIRIC learns interpretable, reusable models, and solves more tasks with fewer environment interactions than all three baselines. On a physical
robot, it learns wind forces and domino masses to solve a manipulation task. Website and code: https://yichao-liang.github.io/empiric
Read the full abstract ▾
All authors Yichao Liang, Amber Li, Dat Nguyen, Emily Bunnapradist, Michelangelo Naim, Sreela Kodali, Matteo Merler, Bowen Li, Kiran Gopinathan, Yiyun Liu, Nikhil Pimpalkhare, Joshua B. Tenenbaum, Adrian Weller, Zenna Tavares, Tom Silver, Kevin Ellis