Computation and Language
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 14 September
Background. Large language models are increasingly used to help write biomedical text but may fabricate references to nonexistent work. How often large language models do so is not well characterized. Methods. We prompted 26 language models from eight developers (2023 to 2026) to supply a missing reference for each of 69 biomedical passages across ten domains. References were classified as verifiable (real paper with a resolving identifier), partial matches (real paper without a resolving identifier), fabricated (no matching indexed paper), or declined (the model refused to supply a reference). A reference was considered correct in every evaluated bibliographic field only when it was verifiable and its journal, year, and listed authors matched those of the cited paper. Results. Fabrication ranged from 10.2% (Claude Opus 4.8, which declined 52.1% of prompts) to 98.4% (Ministral 3B, which produced no verifiable reference). Claude Opus 4.6 and Claude Sonnet 4.5 produced similar proportions of verifiable references (77.6% and 76.6%) but named authors correctly in 78.7% and 28.7% of author-evaluable verifiable references, respectively, and were correct in every evaluated field in 54.6% and 19.9% of responses. GPT-5.5 was correct in every field in 48.1%. Across all models, 55.4% of responses were fabricated and 14.9% were correct in every field. Among the five tested models first released in 2026, the corresponding proportions were 35.3% and 31.8%, respectively. Conclusions. Fabrication remained common, and no model was correct in every evaluated bibliographic field in more than 54.6% of responses. Models that identify real papers may still misstate their metadata, so references produced with model assistance require verification before use.
Cryptography and Security
ActGuard: Pre-execution Action Auditing against Indirect Prompt Injection in LLM Agents
Large language model (LLM) agents interact with external environments through tool invocation, but tool outputs can also expose them to indirect prompt injection (IPI) attacks. Existing defenses mainly rely on prompt hardening, content filtering, pre-generated plans, or permission constraints. These approaches often struggle with complex tasks or over-sanitize external content, making it difficult to balance security and utility. The key challenge is therefore to preserve execution flexibility while precisely identifying and removing the malicious content that actually induces unsafe actions. To address this challenge, we propose ActGuard, a pre-execution action auditing framework. Rather than judging whether external content is inherently suspicious, ActGuard assesses whether it causes the current action to deviate from a locally reasonable expectation. At each step, ActGuard predicts the tools likely to be used by the upcoming action and constructs a local tool prior without constraining the execution trajectory. Before execution, it compares the candidate action against this prior and performs tool-level contrastive analysis and parameter-level evidence localization to identify deviations in tool selection and action parameters. A verifier then examines the localized evidence, masks only spans confirmed as malicious, and regenerates the action from the sanitized context. This design preserves legitimate planning flexibility while minimizing information loss from indiscriminate filtering. We evaluate ActGuard on challenging benchmarks for tool-using agents. Results show that ActGuard reduces attack success rates to a level comparable to state-of-the-art defenses while maintaining task utility close to the no-attack setting, achieving a favorable security-utility trade-off. Our code is publicly available at: https://github.com/binzhwang/ActGuard.
Read the full abstract
Statistical Methodology
Neighbourhood-Based Generalized Dynamic Principal Components for Spatial Functional Data
Regular-grid spatial functional datasets arise naturally in gridded environmental, oceanographic, climate, and remote-sensing applications, where each spatial location is associated with an entire curve. Existing spectral spatial functional principal component analysis methods provide an important frequency-domain approach for such data, but they do not directly target finite-neighbourhood least-squares reconstruction from an estimated latent spatial component field. To address this, we propose Spatial Functional Generalized Dynamic Principal Components (SFGDPC), a reconstruction-based dimension-reduction method for regular-grid spatial functional data. Each function is first represented by basis coefficients, and each coefficient vector is reconstructed from a scalar latent spatial field and its Chebyshev neighbourhood on the rectangular grid. The spatial neighbourhood radius is selected using a Bayesian information criterion (BIC)-type conditional reconstruction criterion. In the local-neighbourhood transfer simulation, SFGDPC reduced mean cumulative normalized mean squared error (NMSE) relative to spatial functional principal component analysis (SFPCA) by approximately 38-53% across the reported component counts and covariance conditions. In the Indian Ocean sea surface temperature (SST) application, one SFGDPC component produced lower whole-grid reconstruction error than both the boundary-safe and high-cap SFPCA benchmarks across all 33 annual fields. The results support SFGDPC as a local reconstruction-based complement to spectral SFPCA for regular-grid spatial functional data.
Read the full abstract
Artificial Intelligence
Converting Sequenced Fuzzy Cognitive Maps to Causal Virtual Worlds with Large Video Generators
We show how users can create and manipulate causal virtual worlds with large-language-model (LLM) and large-video-model agents. The approach uses feedback fuzzy cognitive maps (FCMs) both to model the granular causal structure of the virtual world and to guide its causal evolution. The local causal rules are partial or fuzzy while the FCM's feedback structure produces global equilibria that define causal scenarios. A sequence of dynamical meta-rules of the form “If $\mathcal{A}$ then $\mathcal{B}$" define the causal scenes of the virtual-world video. The if-part causal pattern $\mathcal{A}$ perturbs the FCM's virtual world at the user's or agent's discretion. The FCM's transient feedback dynamics define the meta-rule's causal arrow of implication. The then-part $\mathcal{B}$ is the resulting equilibrium attractor such as a FCM limit cycle or fixed point. Our algorithm extracts these meta-rules from the FCM and guides the LLM agent to write a script based on the FCM meta-rule sequence. The large video generator converts the meta-rule into a video scene in accord with the flow of the dynamics. We applied the agent-based technique to a simple FCM that describes an undersea world of dolphins and sharks. Google's Gemini 3.1 generated the script and Google's Veo 3.1 generated the dolphin-shark video. The approach is general and can scale by mixing larger FCMs and AI agents to produce more immersive virtual worlds.
Read the full abstract
Robotics
Two-Stage Personalized Gait Phase Estimation in Stroke Survivors During Exoskeleton-Assisted Walking: An Offline Feasibility Study
This study evaluated personalized gait phase estimation for stroke survivors using functional inertial measurement unit (IMU) alignment and two-stage sequential adaptation of models pre-trained on healthy gait. The estimator used signals from a thigh-mounted IMU. Heel force-sensitive resistor measurements provided reference phase labels for offline adaptation and evaluation. Stage 1 established a distillation-regularized participant-specific model, and Stage 2 performed conditional refinement using low-rank adaptation. Long Short-Term Memory (LSTM), Temporal Convolutional Network (TCN), and Transformer models were evaluated in five stroke survivors walking with a powered knee exoskeleton using leave-one-subject-out hyperparameter selection and sequential test-then-adapt Stage 2 replay. Relative to the non-adapted baselines, Stage 1+2 reduced the mean participant-wise phase root mean square error by 84.2%, 77.0%, and 60.7%, respectively. The Transformer achieved the lowest final error (2.90 +- 1.13$% of the gait cycle) and heel-strike timing error (23.7 +- 4.5ms). Policy-specific ablations showed that every-cycle updates generally produced the lowest or near-lowest error, whereas conditional updating reduced the update frequency with small accuracy differences. After personalization, alignment produced model-dependent changes in phase error while preserving or improving heel-strike detection and reducing heel-strike timing error for the LSTM and Transformer. Concurrent embedded tests showed that the TCN and Transformer maintained 100-Hz inference during Stage 2 updates without deadline misses, whereas the LSTM missed the 10-ms deadline in 6.6% of inferences. All updates completed within 0.8s. These results support the offline feasibility and embedded computational timing of the proposed framework for exoskeleton-assisted walking.
Read the full abstract
Software Engineering
Cross-Ecosystem Packages As Multilingual: Prevalence, Architecture, and Health
Modern software development increasingly relies on using multiple programming languages. Some software packages are published to multiple package ecosystems such as NPM for JavaScript and PyPI for Python. Little is known about cross-ecosystem packages, especially regarding how they are structured. In this paper, we conduct a large-scale empirical study of over six million packages across six major ecosystems to understand 1) how prevalent cross-ecosystem packages are among all packages, 2) whether there are distinct source code architectural patterns that cross-ecosystem packages use, and 3) whether there are correlations between architectural patterns and project health metrics from GitHub. Results indicate that cross-ecosystem packages constitute a small but important, growing fraction of packages. We identify five distinct architectural patterns. For example, packages that implement code generation from a shared source file or use language bindings are associated with significantly higher community visibility and development activity. Based on our findings, we provide implications for package adopters, maintainers, and researchers. We envision our taxonomy being used for future investigations into several aspects of software development, such as the trade-offs between focusing on one language and translating to other languages using bindings, templating, and wrappers, versus using native code and native functions to support additional languages.
Read the full abstract
Numerical Analysis
efficient matrix-free gauss-newton traveltime tomography on full-metric deformed grids
First-arrival traveltime tomography provides a computationally efficient means of imaging subsurface velocity structure, while boundary-conforming deformed grids allow rugged topography to be represented without abandoning a logically structured mesh. In matrix-free Gauss-Newton inversion, however, the linearized full-metric Eikonal transport and its transpose must be applied repeatedly at a fixed background model. Directional sweeping consequently revisits the same state-dependent dependency structure for every Krylov right-hand side. We recast the frozen tangent transport as a directed graph and distinguish loss of the background-traveltime ordering from genuine algebraic cyclicity. Strongly connected components identify the irreducible part of the transport, whereas the remaining dependencies admit exact scalar substitution after reordering. The condensation graph and local block factorizations are constructed once for each frozen source state and then reused for both primal and transpose applications. Numerical experiments on deformed grids and a three-dimensional tomography problem show that the resulting block-triangular formulation preserves the frozen sensitivity action while substantially reducing the repeated transport cost. By eliminating this dominant inner-iteration cost, the proposed method markedly accelerates Gauss-Newton computation and makes full-metric matrix-free inversion on deformed grids practical at much lower cost.
Read the full abstract
Analysis of PDEs
Variational properties of Perron's extremal solutions in the Bernoulli one-phase problem
We study the Perron extremal solutions of the stationary one-phase Bernoulli problem. These solutions are important in several applications, but not much is known about the regularity of their free boundaries due to their non-variational construction. We show that, in fact, Perron extremal solutions retain some variational features; specifically, they are solutions in the sense of inner variations. We establish this property by showing that extremal solutions arise as infinite-time limits of monotone solutions of the parabolic Bernoulli problem. As a consequence, we are able to describe the structure of singularities for Perron extremal solutions in two dimensions: largest subsolutions have smooth free boundaries, while smallest supersolutions have free boundary points that are either smooth, or whose blow-up limits are two-plane wedges with equal slope.
Read the full abstract
Systems and Control
Topology-Aware Congestion Pricing: Demand Robust Routing using the Forman-Ricci Curvature
Congestion pricing is widely used to improve transportation network efficiency where individual decentralized route choices can lead to system-level inefficiencies. Pigouvian tolls achieve the system optimum but depend on demand and can lose effectiveness under demand shifts, particularly on structurally critical bottleneck edges. We propose a topology-aware regularization of the Beckmann potential governing the (Wardrop) equilibrium by augmenting precomputed Pigouvian tolls with an offline structural penalty derived from Forman–Ricci curvature (FRC), which captures bottleneck structures. We prove that the resulting equilibrium is well-defined, that stronger regularization reduces flow on penalized edges, and that efficiency loss depends on how closely the structural penalty approximates optimal Pigouvian tolls. Experiments on eight real-world networks under targeted and uniform demand perturbations show reduced bottleneck flow with near-optimal efficiency. FRC achieves reductions comparable to edge betweenness at substantially lower computational cost, making it a practical, robust complement to classical congestion pricing.
Read the full abstract
General Relativity and Quantum Cosmology
Quasinormal modes of Reissner-Nordström black hole from bound state spectrum
We apply a newly proposed bound state method to revisit the computation of gravitational and electromagnetic quasinormal modes (QNMs) for Reissner-Nordström black holes (RNBHs). It is found that the method yields QNM frequencies of high accuracy for low-lying modes with overtones $n\leq3$. The accuracy degrades for higher overtones, however, a homotopy deformation to the potential enables us to compute more high overtone modes reliably. It is known that the continued fraction method for computing QNMs of extremal RNBHs differs significantly from that used for nonextremal ones. In contrast, the bound state method demonstrates advantage of simplicity: one only need to account for the definition of tortoise coordinate in the extremal case, and the rest of the computational procedure remains exactly the same as that for the nonextremal case.
Read the full abstract
Information Retrieval
LazFormer: Scaling Transformers for Industrial Recommendation via Transferable Generative Pre-training
Transformers have shown promising performance in LLMs due to their outstanding scalability, several studies have investigated the scalability of Transformers for industrial recommendation. They typically rely on a single ranking model to optimize both sparse and dense parameters from scratch, resulting in substantial computational resource consumption and slow convergence. Fortunately, the pre-training models offer an effective solution to the above issues by providing favorable initialization of both sparse and dense parameters for the subsequent ranking. However, they still face two major limitations: (1) Since the input features used in pre-training and ranking are usually inconsistent, directly transferring dense parameters from pre-training to ranking may lead to negative transfer. (2) Multi-epoch training during the ranking process may result in the overfitting of sparse parameters, while freezing the sparse parameters limits their adaptability to the ranking objectives. To this end, we propose a Scaling Transformer for Industrial Recommendation via Transferable Generative Pre-training, termed LazFormer. Specifically, we first present a generative pre-training module to autoregressively generate sequential features, providing favorable initialization of both sparse and dense parameters for the subsequent ranking. To solve the negative transfer of dense parameters, we propose a transferable residual adapter that injects additional ranking-specific features into ranking in a residual manner. Moreover, a request-aware ranking module integrates long-sequence compression, hybrid sparse attention, and a request-aware paradigm to efficiently model users' long sequences. Besides, we further propose an asymmetric multi-epoch training strategy that resets sparse parameters while continuously accumulating dense parameters across epochs, alleviating the overfitting of sparse parameters.
Read the full abstract
Machine Learning
LiftGCN: Efficient Energy-Preserving Graph Learning via Joukowski Spectral Lifting for Finite Element Stress Prediction
Finite element stress fields often exhibit strong local non-smoothness, where stress concentrations near holes, notches, and loading regions induce sharp spatial gradients and high-frequency graph components. Although graph neural networks naturally operate on irregular finite element meshes, conventional message passing is inherently smoothing and progressively attenuates such high-frequency information. Unitary propagation alleviates this problem by preserving spectral magnitudes, but typically relies on matrix functions and high-order approximations with $O(Ked)$ propagation complexity. We propose LiftGCN, an efficient spectrally stable graph network based on Joukowski spectral lifting. LiftGCN maps the real spectrum of a normalized graph operator onto the unit circle through the Joukowski relation and realizes the resulting spectral transformation as a simple second-order recurrence, avoiding matrix exponentials, eigendecomposition, and high-order polynomial truncation. We show that the linear Joukowski backbone has unit-modulus characteristic roots and admits an energy-preserving structure under a positive-definite metric, preventing exponential attenuation of graph-frequency components with depth. Each layer requires only one sparse neighborhood aggregation, yielding $O(ed)$ propagation complexity, while lightweight local nonlinear residuals provide expressive feature transformations. Experiments on finite element stress prediction demonstrate that LiftGCN achieves competitive overall accuracy while improving reconstruction of stress concentrations and local high-gradient structures with substantially reduced computational cost. Our code is available at https://github.com/ChenZeng001/LiftGCN.
Read the full abstract
Artificial Intelligence
MemRiskBench: Trace-Aware Risk-Preserving Evaluation for Long-Horizon LLM Agents
Long-horizon LLM agents accumulate memory across sessions, creating sparse but high-impact risks: stale facts, conflicting updates, cross-user leakage, revoked-memory reuse, and constraint decay. Standard aggregate scores hide per-risk failure rates–a model achieving 78% average accuracy may still leak data in 4% of episodes–and benchmark compression preferentially discards the rare high-severity events that distinguish a mostly-working model from one that occasionally causes harm. We present MemRiskBench. The primary contribution is a five-category risk taxonomy (plus one documented, unscored category) operationalized by deterministic trace grounded checks, instantiated as a 120-episode scripted benchmark with full trace logging and no LLM-as-judge on the pass/fail path, evaluated on five locally run quantized instruction-tuned models. Second, a risk-preserving subset selector: a coverage-constrained greedy selector on deterministic trace-derived features that retains full ranking (Spearman rho = 0.975, deterministic; CI collapses to a point estimate with zero bootstrap variance), risk coverage (1.0), and high-risk model detection (1.0) at a 20% subset size, reducing compute 5x. Unlike ranking-only subset selectors, this selector additionally preserves risk-type coverage and high-risk detection using trace-grounded deterministic features that do not require an LLM judge. All episodes, traces, the scoring implementation, and the selector are released to support reproducible evaluation and risk assessment of deployed LLM agents
Read the full abstract
Probability
The threshold for online balancing of i.i.d. binary vectors
Consider the task of online vector balancing for stochastic arrivals $X_1,\ldots,{X_T}$, where the $X_i$ are independent uniformly random $d$–sparse binary vectors in $\{0,1\}^n$. This is a random analogue of the online Beck–Fiala problem. We show that uniformly for $2\le d\le n/2$ and $T = \Theta(n)$, the optimal online prefix discrepancy $\max\limits_{t\leq T}\left\|\sum_{i=1}^t\sigma_i X_i\right\|_\infty$ is of order \[ \Theta\big(\max\{\sqrt d,\log\log n\}\big). \] The upper bound is achieved by an efficient online algorithm. Thus, for $d\le(\log\log n)^2$, the optimal discrepancy is $\Theta(\log\log n)$ and is independent of the sparsity up to constant factors, whereas above this scale it is $\Theta(\sqrt d)$, matching the order of the offline discrepancy. This identifies the threshold at which sparsity begins to govern the online discrepancy of the random Beck–Fiala model.
Read the full abstract
Signal Processing
Discrete Antenna Positioning and Beamforming Design for RIS-Assisted MA Secure ISAC Systems
This paper investigates a reconfigurable intelligent surface (RIS)-assisted movable antenna (MA) secure integrated sensing and communication (ISAC) system. In this architecture, the RIS establishes indirect transmission links to provide communication services for multiple legitimate users, while the high spatial diversity gain of MA is leveraged to enhance system security. Then, we formulate an optimization problem to maximize the system total secrecy rate by jointly optimizing the MA position selection, active beamforming design for base station and passive beamforming design for RIS. The problem also accounts for practical constraints including transmit power budget, sensing beampattern mean square error (MSE), RIS unit-modulus constraint. However, it is challenging to solve this problem due to its non-convexity and strong coupling of the optimization variables. Consequently, we propose an alternating optimization (AO) framework, employing techniques including discrete binary particle swarm optimization (BPSO), successive convex approximation (SCA) and difference-of-convex (DC) programming to transform the optimization problem into convex subproblems. Based on the solution above, the convex sub-problems are solved iteratively until convergence is achieved. Numerical results demonstrate that the proposed algorithm outperforms other baseline algorithms in terms of secure communication performance.
Read the full abstract
Computer Vision
PhysBrain 1.5: From Vision-Language Models to Physical Foundation Models
We present PhysBrain 1.5, a unified model for understanding physical environments, generating actions, and predicting future states. Motivated by the physical loop of observation, interaction, and environmental change, we bring these capabilities into a common learning framework. Starting from a general vision–language model, we encode language responses, end-effector motion, and dense visual targets as discrete sequences and jointly optimize them with autoregressive next-token prediction. Pre-training draws its embodied supervision entirely from human interaction videos, using task-centered episodes to pair semantic and spatial context with recovered motion and subsequent observations. We then adapt the model through supervised fine-tuning on a mixture of human demonstrations, robot trajectories, and simulated experience. Across 28 embodied understanding benchmarks, our 8B model achieves an average score of 72.5, setting a new open-source state of the art and performing on par with leading proprietary models such as GPT-6-Astra and Gemini 3.6 Flash. It achieves the best open-source results on 14 benchmarks while retaining general multimodal capabilities. Beyond these understanding evaluations, qualitative examples show the model's ability to produce end-effector trajectories and predict future scenes through spatially aligned RGB, depth, and robot-mask outputs.
Read the full abstract
Machine Learning
Learning to Solve Stochastic Controls with Unknown Drifts and Running Rewards: Theory, Algorithms and Convergence
We study continuous-time and possibly high-dimensional stochastic control problems where drift coefficients and running reward functions are unknown. Due to these missing model primitives, we take the exploratory, reinforcement learning (RL) framework of Wang, Zariphopoulou, and Zhou(2020) with relaxed controls and entropy regularization. The objective is to develop theoretically grounded, efficient and scalable RL algorithms to learn both the optimal value functions (which also solve the exploratory HJB equation) and optimal exploratory feedback control policies. When the diffusion coefficients do not contain control, we employ probabilistic representations of both the optimal value function and its gradient based on an auxiliary state process depending only on the diffusion part of the original dynamics. With a delicate analysis on some properly defined mappings and their fixed points, this leads to the introduction of our policy iteration algorithms and their convergence. We demonstrate the performance of our algorithms through various numerical examples. Finally, we study a special control-dependent diffusion case where probability representation of the Hessian is called for.
Read the full abstract
Graphics
Opacity Is Not Just Opacity
Web graphics travel with content across pages and themes, where changing backgrounds can require recoloring and maintenance. Opacity already makes a fixed object's appearance depend on its background, yet is usually understood only as how much the object obscures it. In fact, opacity controls the scaling of the object-background color difference; transparency is only one effect of this relationship. Zero places the output at the background and one at the source color, but difference scaling need not stop at either position. We retain the compositing expression and extend the coefficient domain from $[0,1]$ to the real numbers: negative values reverse the difference, whereas values above one expand it in the same direction. We focus on same-direction expansion for reusing Web graphics across backgrounds. Each object carries a fixed source color and coefficient, while the actual background determines the enhancement direction. Background-adaptive contrast enhancement thus becomes part of the object's compositing properties, reducing the design and maintenance of separate color variants. The implementation reuses the original equation without increasing the per-pixel arithmetic operation count within the same pipeline. Enumerating all 8-bit sRGB source colors on 16 predefined light and dark canvases, a fixed $\alpha=1.1$ increases the contrast ratio in 99.8145% of combinations. Without changing source colors, 4.8346% of all combinations newly reach the $3:1$ contrast threshold. Output validation and timing across three browsers demonstrate implementation in the same WebGL pipeline, with no sustained additional runtime observed.
Read the full abstract
Artificial Intelligence
Towards a knowledge-enhanced single-cell foundation model
Single-cell foundation models (scFMs) increasingly rely on large-scale transcriptomic pretraining, yet expanding pretraining data can yield diminishing gains while substantially increasing computational cost. Our data scaling analyses showed that incorporating biological knowledge, including cell-level text annotation and gene-level regulatory information, provided additional scaling dimension than simply increasing data size. Motivated by this observation, we present scKITE, a simple yet effective scFM that integrates cell-annotation and gene-regulatory supervision into a shared transcriptomic Transformer encoder through lightweight auxiliary decoders. These decoders are used only during pretraining and subsequently discarded, yielding a general-purpose encoder enriched with biological knowledge for downstream applications. With only 179,067 pretraining samples, i.e., less than 0.5% of those used by previous strong scFMs, scKITE outperformed these models across diverse downstream tasks, highlighting knowledge-enhanced pretraining as a promising paradigm for biologically grounded scFMs.
Read the full abstract
Computation and Language
Online Language Adaptive Sampling for Better Distributed Cross-lingual Gains
Realignment is a promising approach for improving the cross-lingual transfer ability of multilingual language models, particularly for extremely low-resource languages (LRLs). However, existing realignment methods rely on uniform and random sampling of parallel sentences across languages, which may be suboptimal under limited batch sizes. In practice, models may benefit from seeing certain languages more frequently, especially those that are poorly aligned, and the optimal distribution can evolve throughout training. In this work, we propose a simple yet effective adaptive sampling strategy that assigns trainable sampling probabilities to each language. Languages that contribute more to the realignment loss are sampled more frequently in subsequent batches, and the optimal distribution can evolve throughout training. Our method employs an inner-outer optimization loop with a small overhead, leading to consistent performance improvements and, more importantly, distributing the gains across languages. We observed a $+0.67$ average performance increase on all tasks with XLM-R, and $+0.60$ with Gemma 2 9B compared with uniform realignment. Furthermore, our method is robust across different models. Code available at https://github.com/felixgaschi/multilingual-alignment-and-transfer.
Read the full abstract
Machine Learning
HiGFRL: Hierarchical Graph Fusion-Driven Reinforcement Learning for Dependency-Aware Task Scheduling in Heterogeneous Cloud
Online scheduling of dependency-aware tasks in heterogeneous cloud clusters is a fundamental yet challenging problem due to the complex interplay between DAG topologies and multi-dimensional resource constraints. While DRL has shown promise, existing GNN-based approaches often struggle to efficiently model high-order topological dependencies and suffer from loose coupling between task and resource states, leading to myopic scheduling decisions. To address these limitations, we propose HiGFRL, a Hierarchical Graph Fusion-Driven Reinforcement Learning framework. HiGFRL constructs a novel three-level state representation comprising a Static Hypergraph, a Dynamic Global Graph, and a Local Bipartite Graph to explicitly model the interplay between task dependencies and real-time cluster dynamics. Specifically, we design a fusion-driven dual-network architecture to optimize RL decision-making, where a Context Fusion Allocator integrates local bipartite matching features with fused global context to execute precise task-to-node allocation, and a Global State Evaluator leverages the global dynamic graph representation to accurately estimate expected long-term cumulative reward. Furthermore, we incorporate a topology-prior-guided hybrid reward mechanism that distills static topological priors into the learning process to accelerate convergence. Extensive experiments using real-world Alibaba cluster traces demonstrate that HiGFRL significantly outperforms heuristics and DRL baselines. Specifically, in challenging large-scale high-load scenarios, HiGFRL reduces the Makespan by up to 32.55%, and optimizes the average task flow time and average task wait time by 13.58% and 13.79%, respectively. Experimental results confirm that HiGFRL not only significantly improves cluster throughput but also ensures superior QoS by substantially reducing queuing delays. Code Release:https://github.com/igeng/HiGFRL.
Read the full abstract
Computation and Language
A Corpus-Aligned Uthmani-to-Standard Quranic Word Mapping and a Deterministic Recitation Validator
Quranic text is distributed in two orthographic forms that are byte-level distinct: the Uthmani script used in every printed mushaf, and the Standard (Imla'i) Arabic form that every mainstream Arabic NLP tool is built for. The gap is concentrated in one Unicode character, U+0670 (superscript alef), which appears in some of the most frequently recited words in the Quran and is silently mishandled by general-purpose Arabic normalizers. We release a 2,290-pair, corpus-aligned Uthmani-to-Standard word mapping constructed by aligning the complete 6,236-verse Quran across both orthographic forms, together with a seven-step text normalization pipeline built on it. Normalizing both forms of all 6,236 verses through that pipeline yields identical strings for 90.9% of verses, and we characterize the residual divergence rather than assert that it is closed. On top of the normalized text, we build a deterministic, LLM-free Quranic recitation validator using a four-layer verse-matching search (exact, morphological, relaxed, fuzzy) and word-error-rate-graded feedback across five severity tiers. The validator scores 98.4% (122/124) on a 124-case suite emitted by the released test harness, and both failures share one mechanism: a single substitution error can make a different verse an exact match. A full-corpus census additionally quantifies an inherent text-only ambiguity affecting 16.5% of verses, and on 34 recitation transcripts drawn from a deployed Arabic ASR system the validator identifies the correct verse in every case. We release the mapping, the script that builds it, the validator, and the evaluation harness under open licenses; every number in this paper except the deployment measurement, whose transcripts are not ours to publish, is reproduced by running them.
Read the full abstract
Cosmology and Nongalactic Astrophysics
Forecasting Coupled Dark Energy Parameters with the One-Loop Galaxy Power Spectrum
We forecast constraints on the parameters of the coupled dark energy model using DESI and Euclid data with the one-loop galaxy power spectrum. We investigate the distinguishability of our model from the zero-coupling scenario at the $1\sigma$ level and explore how the parameter constraints depend on the fiducial coupling $\beta_{\rm fid}$, the fixed potential slope parameter $\nu$, the maximum wavenumber $k_{\max}$, and different prior choices. We find that the inclusion of mildly non-linear scales improves the constraints on the coupling by roughly a factor of five. Then, we address the question of which is the minimum value of $\beta$ that can be distinguished from zero at $1\sigma$. We find that for our reference case $k_{\rm max}=0.2 h/$Mpc, $\beta=0.15$ lies roughly $1\sigma$ above zero. In the most optimistic case with $k_{\rm max}=0.3 h/$Mpc and including a Planck prior on $\Omega_{m0}$, this value can be reduced to $0.05$. These values are substantially larger than the current constraints on $\beta$, but the latter have been obtained assuming $\beta$ to be constant from at least the decoupling epoch to today, while we only employ late-time data. We conclude therefore that only models that allow for time-varying couplings can be detected with late-time datasets.
Read the full abstract
Computer Vision
MoVT: Video-Augmented Motion Tokenizer for Text-to-Motion Generation
Text-driven 3D human motion generation models face significant challenges in responding to diverse and unconstrained textual prompts, primarily due to the limited availability of 3D motion training data. To address this, we introduce MoVT, a novel framework that effectively leverages the extensive range of human action videos to enhance text-to-motion generation. At the core of our approach is the cross-modal augmented motion tokenizer, which projects discrete 3D motion tokens into the 2D domain. This projection allows us to enrich the motion codebook with complex, real-world motion patterns derived from videos. The enriched discrete tokens are then mapped back to the 3D domain, resulting in aligned 3D and 2D codebooks with an enhanced capacity to represent intricate motions. These enhanced codebooks are integrated into a generative masked transformer, which predicts masked motion token indices in a modality-agnostic manner. This enables the use of text-index pairs, generated from the 2D codebook and annotated motion videos, to further enhance the generator. Extensive empirical evaluations show that MoVT performs favorably against prior state-of-the-art methods across multiple key metrics.
Read the full abstract
Quantum Physics
Exact Breit-Rabi formulae for the nitrogen-vacancy center in diamond
The hyperfine structure that the nitrogen-vacancy (NV) center in diamond has with its nitrogen nucleus is of immense importance for quantum sensing and quantum computing. Current methods of calculating the NV hyperfine energy eigenvalues and eigenstates with respect to an axial magnetic field - i.e., calculating the NV Breit-Rabi formula - resort to numerical methods and analytic approximations. Here we provide an exact, closed-form expression to the NV hyperfine energy eigenvalues and eigenstates for both the nitrogen-14 and -15 cases.
Read the full abstract
Computation and Language
Can We Triage LLM Translation Errors in Classical Texts Without Human References? Source Novelty, GEMBA Scoring, and Budgeted Review through Pali-to-English Translation
As large language models become capable translators of classical texts, a key challenge is deciding which outputs need expert review when no human reference exists. This study tests reference-free error triage through Pali-to-English translation. Three LLMs translated 15,493 passages. Five signals were compared: source novelty, source-candidate embedding distance, peer-translation disagreement, English-to-Pali backtranslation, and no-reference GEMBA scoring. Signals were calibrated on a 3,000-item reference-informed LLM-adjudicated sample and checked against a 500-item author-adjudicated anchor. Human references supported calibration and validation only; they were never used to compute the risk signals. Source novelty was a useful source-side risk prior but not a per-candidate error detector. Peer disagreement and backtranslation provided secondary signal. The strongest method was no-reference GEMBA scoring by a panel of models generally regarded as stronger than the translators: reviewing the top 10% by GEMBA risk captured 81.6% of panel-major errors in the calibration set. GEMBA also remained the best reference-free signal against the author anchor. A same-tier panel, with self-scoring excluded, remained useful but performed worse, indicating that evaluator strength matters beyond the prompt alone. A budgeted workflow is proposed, combining source novelty, peer disagreement, and stronger candidate-aware judging to allocate human review. Transfer to other classical languages, including Latin, Ancient Greek, and Sanskrit, remains to be tested.
Read the full abstract
Artificial Intelligence
Geometric Flow enhanced Graph Coarsening
Recently, researchers have proposed a graph pooling operation, akin to the pooling process in conventional convolutional neural networks (CNN), aimed at reducing the computation cost of Graph convolutional neural networks (GCNNs). While most GCNN-based methods treat graph pooling as a node clustering problem and propose learning a cluster assignment matrix, existing clustering-based pooling methods tend to focus solely on the rough topology information of graphs, neglecting the exploitation of higher-order mutual connections among neighbors. In terms of message passing on graph, the ease of information passing on edges reflects the closeness between neighboring nodes, which significantly relies on the interconnectivity among neighbors. In this study, we address this gap by considering such local connection information and introducing a novel graph pooling method named RicciPool. We introduce discrete graph curvature, particularly Ollivier-Ricci curvature, as a measure of higher-order connectivity around an edge. Subsequently, we construct an Ollivier-Ricci flow formula to reweigh edge weights, leveraging the crucial information provided by Ricci curvature, particularly vital for extracting clusters in graphs. Building upon this foundation, we utilize the spectral clustering technique to learn a new cluster assignment matrix. Experimental results on multiple bioinformatics protein datasets and social networks underscore the effectiveness of our proposed method.
Read the full abstract
Operator Algebras
Discrete $C^*$-Frobenius Algebras and Infinite-Index Extensions in Algebraic Quantum Field Theories
We introduce discrete $C^*$-Frobenius algebras as categorical data for constructing infinite-index extensions of Möbius covariant nets. The underlying representation is a countable direct sum of simple dualized modules, while the direct sum itself is not necessarily dualizable. Accordingly, rather than requiring a single bounded multiplication morphism, we work with compatible families of bounded left and right multiplication morphisms. Using Bin Gui's theory of categorical extensions, we associate to every discrete $C^*$-Frobenius algebra two (generally non-local) Möbius covariant extensions generated by left and right charged fields, respectively. If the discrete $C^*$-Frobenius algebra is commutative, the two extensions coincide and yield a local Möbius covariant extension. As applications, for every $C^*$-tensor category generated by an invertible object with trivial self-braiding such that its fusion product rule is given by $\mathbb{Z}$, we construct two classes of examples: the simple current extension by $\mathbb{Z}$ and the discrete Longo-Rehren construction.
Read the full abstract
Quantum Physics
Universal Defect Statistics in Reverse Quenches
We extend the generalized Kibble-Zurek mechanism (gKZM) to reverse quench protocols using a stochastic description of defect formation finding that the mean defect density after a reverse quench equals twice the forward quench defect-number variance per unit volume. This relation, universally valid within the stochastic framework of the gKZM, makes the forward quench variance accessible through a reverse quench defect density measurement. We verify the equality analytically and numerically in the uniform one-dimensional transverse-field Ising model, and numerically in its bond-disordered variant, including a regime where defect scaling departs from a pure power-law in quench time. Our result also suggests an experimental route to probing quantum critical behavior: reverse quenches that begin and end in the paramagnetic phase may allow to reveal critical exponents using global magnetic-field control and magnetization measurements, without resolving individual defects.
Read the full abstract
Machine Learning
High-Probability Nash Regret for Decentralized Learning in Markov $\alpha$-Potential Games: Episodic and Fully Online Asynchronous Algorithms with Applications to Markov Congestion Games
We study decentralized learning of Nash equilibria (NE) in infinite-horizon discounted Markov games under bandit feedback, focusing on Markov $\alpha$-potential games. We develop KL-projected natural policy gradient (NPG) algorithms in two settings: an episodic setting with frozen policies during sampling and a fully online setting in which players receive a single realized cost sample per time step and update their policies asynchronously along a continuing trajectory. We establish finite-time high-probability NE regret bounds of order $\widetilde O(T^{-1/4})$ and $\widetilde O(T^{-2/15})$ for the episodic and fully online settings, respectively, up to fixed approximation terms. Crucially, our bounds eliminate the distribution-mismatch coefficient, which can scale prohibitively with the size of the state space, while accommodating potential approximation, estimation-oracle bias, and transition sensitivity. We further identify a state-wise potential structure that yields sharper guarantees with additive dependence on the potential approximation error $\alpha$. We specialize the framework to independent-resource Markov congestion games (IMCGs), establish their approximate-potential and transition-sensitivity properties, and construct decentralized estimation oracles from realized costs. As an application, we introduce strategic online job scheduling on stochastic machines and obtain a scalable decentralized algorithm for learning stable dispatching policies. Overall, our results provide the first finite-time high-probability NE regret guarantees for fully online asynchronous decentralized learning in Markov $\alpha$-potential games, remove distribution-mismatch coefficients from the regret bounds, accommodate fixed estimation-oracle bias, and provide scalable decentralized learning with finite-time guarantees for IMCGs.