Papers

Batched simulators for autonomous driving have recently enabled training reinforcement learning (RL) agents at scale, encompassing thousands of traffic scenarios and billions of interactions within a matter of days. Although such high-throughput feeds RL algorithms faster than ever, their sample-efficiency has not kept pace: As the standard training scheme, domain randomization uniformly samples scenarios, thereby consuming a vast number of interactions on cases that contribute little to learning. Curriculum learning offers a remedy by adaptively prioritizing scenarios that matter most to policy improvement. We present CL4AD, the first integration of curriculum learning into batched autonomous driving simulators by framing scenario selection as an unsupervised environment design problem. We introduce utility functions that shape curricula based on success rates and the realism of the agent's behavior, in addition to existing regret-estimation functions. Large-scale experiments in GPUDRIVE demonstrate that curriculum learning achieves a 99% success rate a billion steps earlier than domain randomization, reducing wall-clock time by 77%, and outperforms heuristic curricula with static and dynamic attributes, with only one exception at the largest scale. An ablation under limited compute shows that curriculum learning improves sample efficiency by 67%. We also investigate how utility functions behave at scale, and how prioritized scenarios evolve during training. We release an implementation of CLForAD in GPUDRIVE.

Topic profiles represent publications, authors, and other scholarly entities as probability distributions over a fixed set of topics, but flat total variation treats every pair of distinct pure-topic profiles as maximally separated and therefore ignores taxonomic proximity. From a rooted weighted taxonomy, we derive a cardinality-normalized linear operator that maps the leaf topics to points in the original probability simplex and exactly realizes a normalized lowest-common-ancestor ultrametric under total variation. The operator is doubly stochastic and positive definite; within the class of nonnegative edge-cluster Gram operators, its normalization is uniquely determined on the reduced branching tree. Applying the same invertible operator to arbitrary topic mixtures yields a nondegenerate hierarchy-aware metric that contracts flat total variation, differs from the tree-Wasserstein distance on mixtures, and can be evaluated in O(|V|+L) time and memory without forming the dense matrix. In a frozen OpenAlex taxonomy with 4,516 terminal Topics, raw dissimilarities between Topic texts showed consistent ordinal alignment with taxonomic proximity, while only 3 of 253 calibrated internal nodes required monotonic correction. Encoder choice nevertheless affected individual height estimates. The framework exactly realizes a supplied weighted hierarchy; text is used only to initialize its node heights, and distances between scholarly topic profiles are then computed in the induced geometry.

Enterprise agents are expected to deliver correct and consistent outcomes while executing complex procedures, adhering to governing policies, and operating within practical resource budgets. We present Stage, an integrated execution framework combining typed task nodes, shared execution memory, document tools, and deterministic orchestration. We evaluate Stage on SOP-Bench Referral Abuse, the Retail and Telecom domains of $\tau^2$-bench, and two proprietary banking workflows, Smart Dispute and Paystub Fraud Detection. Using the same underlying model in each comparison, Stage improves single-run success and repeated-run reliability over the baseline in all 18 benchmark–model pairs. Success across all three runs (\(\mathrm{Pass}^{3}\)) increases by 7.5–66.0 percentage points. Fixed-graph ablations demonstrate the contributions of localized policy context and coordinator-owned routing. Further analysis of Smart Dispute shows improved coverage of required operations across procedure lengths. These findings show that integrating local reasoning, persistent state, and procedural control can improve reliable enterprise workflow completion without changing the underlying model.

Suppose $X,Y$ are independent random variables with values in a compact abelian group $(G,+)$. We examine the following two entropy power-type inequalities: $h(X+Y)\geq \frac{1}{2}h(X)+\frac{1}{2}h(Y)$ and $h(X+Y)\geq \max\{h(X),h(Y)\}$, where the entropy $h(Z)$ of a $G$-valued random variable $Z$ is defined in terms of its density with respect to Haar measure on $G$. For groups that are either connected or finite with no nontrivial subgroups, we precisely characterize the cases of equality and establish explicit, quantitative stability estimates in terms of relative entropy for these two inequalities. The main tools are a generalization of an entropic inequality obtained by Green, Manners and Tao (2023) for discrete entropy, and a harmonic-analytic estimate for the chi-squared contraction coefficient in connected compact groups. As an application, we derive exponential convergence rates to the uniform distribution in relative entropy for random walks on connected compact abelian groups.

Autonomous computer-use agents are increasingly applied to long-horizon tasks requiring coordinated application calls, persistent state tracking, and verifier-sensitive writes, yet they remain prone to procedural failures: misreading application state, tool semantics, or task progress. Procedural memory promises more consistent decisions and less redundant exploration, but constructing high-quality memory without model training remains challenging. We introduce CONTRAMEM, a source-flexible, training-free framework for self-evolving procedural memory that treats same-task outcome variation as supervision: differences in correctness, efficiency, recovery, and failure modes expose outcome-relevant procedural distinctions, distilled into a compact bank of app-level Function Cards and task-level Skill Cards that evolves through localized curation rather than append-only accumulation or whole-bank rewriting. On held-out GAIA2/ARE computer-use tasks, CONTRAMEM more than doubles the success rate across the three source-model targets (26.2% to 55.3%), with consistent per-model gains (GPT-5.5: 27.5 to 61.0; Claude Sonnet 4.6: 28.0 to 52.5; DeepSeek V4 Pro: 23.0 to 52.5). The same bank transfers unchanged to the unseen Qwen3.7 Plus (18.5 to 35.5), indicating transferable procedural knowledge rather than model-specific behavior. The same construction carries over unchanged to AppWorld, beating both no memory and its own single-source self-memory variant for all three mid-tier agents on both public test splits. Under a matched trajectory budget, heterogeneous multi-model trajectories yield stronger memory than self- or same-model multi-rollout memory: the margin comes from contrastive behavioral diversity, not stronger source agents or more sampling.

Domain shift across imaging modalities and acquisition sites remains a significant barrier to the clinical deployment of segmentation models. Source-free unsupervised domain adaptation (SFUDA) addresses this by adapting a pretrained model to an unlabeled target domain without requiring access to sensitive source data. We introduce a novel SFUDA framework built on Symmetrical Flow Matching, a unified generative model that segments an input image and synthesizes a source-like image from a mask within the same learned flow. By initializing inference from a domain-agnostic Gaussian origin, the model preserves structural consistency across domains and grounds predictions in learned anatomy rather than shifted texture statistics. Our pipeline leverages this symmetry to generate reliable pseudo-labels and corresponding source-like synthetic images from unlabeled target data, creating a generative replay buffer that anchors source knowledge during a generative self-training stage that fine-tunes on a joint set of real target and synthetic source-like images. We evaluate on abdominal multi-organ and cardiac segmentation, covering cross-modality MRI<->CT shifts, and multi-site prostate segmentation. Our approach outperforms SFUDA baselines and is competitive with conventional UDA methods.

Evaluating Large Language Model (LLM) code generation quality requires examining not just whether the generated code is correct, but whether it is maintainable, efficient, and stylistically sound, all of which are qualities of direct importance to software engineering practitioners. Existing benchmarks reduce evaluation to a single Pass@k metric, which obscures critical trade-offs between functional correctness and structural quality. A further limitation is the near-exclusive focus on Python, leaving enterprise-relevant ecosystems such as C# and .NET without dedicated evaluation. This paper presents an automated, multi-dimensional evaluation framework for C# code generation, applying it to four state-of-the-art LLMs: GPT, Gemini, Claude, and Grok. We conduct a controlled experiment across 85 algorithmic tasks derived from HumanEval, generating and evaluating 340 solutions in total, in which each solution is assessed across three independent dimensions: functional correctness via automated unit testing, static code quality via Roslyn AST analysis, and runtime efficiency via adversarial BenchmarkDotNet profiling. Our central finding reveals a substantial gap between correctness and quality attributes (Pearson r = 0.075), demonstrating that Pass@k rankings systematically misrepresent the full LLM performance profile in software engineering contexts. We further characterize GPT's bimodal failure behavior.

This article establishes the quasi-optimality of adaptive nonconforming finite element methods for a class of optimal control problems involving Dirac measures governed by the biharmonic equation. The nonconforming Morley finite elements are employed for discretising both the state and adjoint variables. A modified right-hand side through a companion operator that maps Morley finite elements toaconformingspacehelpstoovercomethechallengeinhandlingpointsourcesontheright-hand side. A priori and a posteriori error estimates for the optimal control problems are derived. Further,optimal convergence rates for adaptive finite element methods are established using an axiomatic framework: by proving key properties such as stability, reduction, discrete reliability, and quasi-orthogonality. Numerical experiments for three types of optimal control problems are discussed extensively and they validate the theoretical results.

Recently, machine unlearning, the removal of specific training data influence from a model, has gained increasing attention. In large language models (LLMs), unlearning is particularly challenging due to the ambiguity of inputs and outputs. Con- sequently, rigorous evaluation is critical for assessing both safety and utility, and for driving progress in unlearning meth- ods. We identify two key shortcomings in existing unlearning benchmarks: (1) they do not actively test whether unlearned information can still be forcibly extracted, and (2) they fail to evaluate performance preservation on boundary questions, be- nign queries that are semantically close to the unlearned con- tent. Here we introduce WMDP++, an extension of WMDP that addresses these gaps by incorporating targeted extrac- tion of unlearned information and systematic evaluation on boundary questions. WMDP++ provides a more stringent and informative benchmark for evaluating unlearning in LLMs.

We introduce RS$^3$-Prune, a training-free token-pruning recipe that instantiates as a small set of inference time hooks atop existing video object segmentation (VOS) networks. Modern VOS models have converged on a common, expensive design: an image encoder produces a dense token grid for every frame, and a memory bank accumulates these tokens across all previously processed frames to condition future predictions. As a video grows longer, the resulting token budget governs both per-frame latency and peak GPU memory. Hence these models break on use cases such as — long-form video or real-time deployment on memory-bounded accelerators. In this work we argue that the right axis along which to compress memory-bank VOS is the token budget itself. RS$^3$-Prune operates in two precise locations within an arbitrary memory-bank VOS pipeline: at the boundary between the image encoder and the memory-attention readout, where we restrict the queries that participate in the cross-frame attention to only a small, geometrically informed subset; and at the boundary between the memory encoder and the memory bank, where we restrict which tokens are ever permitted to enter the bank to those that lie within the object's spatial extent. Over various established benchmarks, RS$^3$-Prune delivers up to $38.8\%$ FPS speedup and reduces $13.1\%$ peak memory usage, while preserving a competitive $\mathcal{J}$&$\mathcal{F}$ compared to the unmodified VOS networks.

Private bidding is a process in which participants submit sealed bids, ensuring that their content remains hidden from other bidders during the bidding window. This is essential in competitive environments to ensure a fair and independent evaluation of all proposals. While (public) blockchain enables decentralized bidding and its transparency offers advantages such as public verifiability, without a Trusted Third Party (TTP), current methods struggle to verify whether a bidder is eligible. We propose a framework that leverages Self-Sovereign Identity (SSI) to address these issues by certifying the bid's eligibility and authenticity using the now-established SSI framework. To allow participants to prove they meet the requirements while keeping their bids secret, our model uses Verifiable Credentials (VCs) and cryptographic primitives such as Zero-Knowledge Proofs (ZKPs).

We study one-sided online bipartite matching with recourse. In this setting, one side of a bipartite graph is known in advance, while vertices on the other side arrive online together with their incident edges. After each arrival, the algorithm must maintain a maximum-cardinality matching while minimizing the total number of reallocations, also known as the recourse budget. Despite extensive work, the exact recourse complexity of the problem remains unsettled: the best lower bound is $\Omega(n \log n)$, whereas the best upper bound is $\mathcal{O}(n \log^2 n)$, where $n$ denotes the number of online vertices. Tight upper bounds of $\mathcal{O}(n \log n)$ are known only for restricted graph classes, such as forests. The best known upper bounds are attained by a very simple and natural algorithm SAP, which after each arrival applies a shortest augmenting path, and it is conjectured to be optimal. All known upper bound analyses of this algorithm do not depend on the particular maximum matching maintained by the algorithm. Consequently, they also apply to a more difficult problem, which we call the malicious matching setting: after each arrival, the maintained matching is replaced by a worst-case maximum matching for the next step. This led to the conjecture that the malicious setting still admits an $\mathcal{O}(n \log n)$ recourse bound, in line with the conjectured optimal complexity of the original model. Our main result is an $\Omega(n \log^2 n)$ lower bound for the malicious matching setting, thus disproving the conjecture. Together with the previous upper bound, this settles the asymptotic recourse complexity of the malicious variant of the problem. We complement our lower bound with an upper bound of $\mathcal{O}(n \log n)$ for expander graphs.

Data valuation is a cornerstone of data-centric learning, where prior efforts primarily focus on designing algorithms to classify training samples as either beneficial or detrimental for the learning task. However, leveraging these valuation estimates for subsequent data intervention remains underexplored; conventional approaches typically discard or downweight harmful samples, thereby underutilizing available data resources. In this paper, we present Dynamic Influence-based Valuation and Editing (DIVE), a novel and efficient framework that dynamically estimates sample values at the batch level and transforms detrimental data into beneficial contributions. Rather than altering the raw data, DIVE operates at the optimization level by strategically reversing the gradient directions of harmful samples during training, ensuring seamless integration with standard learning procedures with minimal overhead. Extensive empirical evaluations demonstrate that DIVE consistently improves classification performance, maximizes data efficiency, stabilizes optimization, and effectively generalizes to large language model fine-tuning.

Visual autoregressive (VAR) models offer fast, high-quality text-to-image generation but still fail at attribute binding and spatial relations. Existing VAR test-time methods search sampled trajectories or apply generic logit guidance but do not directly optimize compositional constraints; gradient-based diffusion methods do not transfer to stateful, discrete, multi-resolution VAR sampling. We introduce $\textbf{VISTA}$ ($\textbf{Vi}$sual Autoregressive $\textbf{S}$emantic $\textbf{T}$est-time $\textbf{A}$lignment), the first gradient-based test-time framework for compositional alignment in next-scale autoregressive image generation. Built on Infinity, VISTA optimizes intermediate representations through the frozen transformer without parameter updates or training. It stabilizes optimization across scales and exposes a common interface into which any differentiable constraint on cross-attention can be plugged. Across two benchmarks and model scales, VISTA improves every targeted category, raising the mean by nearly 20% on a 2B backbone and almost 6% on an 8B backbone, especially on spatial relations. In human evaluation, VISTA wins 83% of decisive compositionality comparisons by plurality vote, with no significant preference in perceived quality; independently, ImageReward scores its outputs 20% higher. VISTA enables the 2B model to surpass a backbone four times its size, showing much of the compositional gap is recoverable at test time.

The Hadamard maximal determinant problem seeks an $n\times n$ matrix $X$ with entries in $\{\pm 1\}$, which maximizes the determinant $\vert \det X \vert$ for a given order $n \in \mathbb{N}$. We report matrices attaining new record determinants for orders $n\equiv 3\pmod{4}$ between $103$ and $119$, as well as $n=51$. We provide a proof of optimality within a specific family of circulant–block constructions for $X$. For these orders, we recast the Hadamard problem as a sequence reconstruction problem from a pair of periodic autocorrelation functions, subject to certain arithmetic conditions. We describe several learning–based strategies for reconstruction and provide a comparison with classical simulated annealing–style approaches as a benchmark.

A long-video answer is evidence-supported only when the frames decoded from the video cover every event the answer depends on. Existing evaluations score final-answer correctness or predicted evidence intervals, but the frames a method decodes before answering are rarely audited, so correct answers can still rest on incomplete observation. We introduce VES-Bench, a 600-question benchmark of Temporal Ordering and Event Counting items over 348 public long videos. Each item carries a jointly necessary set of evidence intervals, letting us audit at three strictness levels whether a method's decoded frames cover every one of them. We also propose TRACE, a training-free agent that grounds answers in raw visual clips, builds an evidence bundle round by round, and stops only when the answer stabilises as the bundle grows and a final pass over the same clips returns the same answer. Under a same-backbone audit, TRACE answers 50.7% of questions correctly with at least two decoded frames inside every evidence interval, at 98.7 frames per question: over 10 points above uniform decoding at 128 frames (40.2%), and within 2.6 points of uniform decoding at 256 frames at 0.39x its frame cost, while reaching the highest answer accuracy in the audit (63.5%). TRACE also stays competitive on Video-MME (86.1), LVBench (75.6), and LongVideoBench (75.1).

Reliable early disruption prediction is critical for the safe operation and real-time control of tokamaks. However, machine learning based prediction frameworks have predominantly targeted medium and long pulse devices, with comparatively limited attention given to short pulse tokamaks where available warning time is inherently constrained. In this work, an interpretable machine learning framework is developed for feature engineering and early prediction of disruptions in the ADITYA using the initial plasma evolution information, prior to the activation of the negative converter of the ohmic transformer power supply. Statistical descriptors comprising the mean, variance, skewness, kurtosis and wavelet energy entropy are extracted from routinely available plasma diagnostics over different operation time windows. Decision tree based feature selection is employed to identify physically meaningful disruption precursors and to reduce feature dimensionality. These selected features are used to train a random forest classifier. The proposed framework achieves stable predictive performance across different analysis windows, with a maximum ROC-AUC of 0.87 for 0-35 ms and 0-40 ms windows. Comparable and in some cases improved, performance is obtained using the reduced feature set, demonstrating that the selected statistical descriptors retain the essential information required for disruption prediction. The proposed methodology provides an interpretable and computationally efficient framework for real time disruption prediction in short pulse tokamaks and establishes that carefully engineered statistical descriptors can effectively replace raw time series inputs for early disruption prediction, thereby offering a practical pathway toward real time plasma control in short pulse tokamaks similar to ADITYA and ADITYA-U.

Neural networks can learn algebraic operations from finite examples, but it remains unclear whether this ability transfers across mathematically equivalent representations of the same operation. We study this question through multiplication in finite fields under changes of basis. The Galois action organizes basis representations into orbits, and bases in the same orbit induce the same coordinate multiplication map. This structure allows us to separate learning multiplication from transferring it to basis representations that are not used for training. We examine several ways of providing or recovering the relevant orbit structure, including invariant labels, basis matrices, orbit recognition, and algebraic decomposition. Our main approach trains a model to predict the Galois action between basis representations. Repeated applications of the learned transformation are then used to construct a canonical representative for each orbit, which supports multiplication on held-out bases through exact canonical matching. This provides a concrete mechanism for converting a learned algebraic symmetry into an invariant representation that can be used for transfer.

Autonomous multi-agent systems nowadays act in finance, software supply chains, and security operations. Already, the first largely AI-orchestrated intrusion campaigns have been reported. Yet, when such a system causes harm, no method can robustly establish what happened, what caused it, or who is accountable. This is because provenance forensics works at the wrong abstraction, formal causality assumes the causal model, and agent auditing trusts self-recording. The target failure mode is, thus, attribution laundering, i.e., spreading an act across redundant agents until none is a but-for cause. Worse, the record is produced by the suspects, which comprises the assumption adopted throughout this work. Agents may therefore anticipate the investigation and the part of logging infrastructure may itself collude. In this paper, HANSARD is proposed, a reference architecture treating accountability as a life-cycle property. First, a readiness profile sealed before operation bounds what later findings may claim. Second, capturing at five choke points beyond the agents' reach makes omissions detectable, not only tampering. Third, a typed PROV-DM-aligned causal graph accrues as the system runs, and three indicators read it live to gate oversight without adjudicating. Fourth, post-incident replay yields contingent effects under the modified Halpern-Pearl definition, together with a compensation-set size. Finally, a synergy residual measures harm due to the combination rather than to individuals, making laundering visible. Cause, responsibility and accountability are then reported separately, each capped by an evidentiary tier, while a future research agenda is also provided.

Agent benchmarks often evaluate only final answers even when agents run on stateful runtimes. We argue this under-specifies what is being evaluated: the proper unit is a declared model-plus-runtime configuration whose failures can occur in evidence acquisition, runtime routing, safety boundaries, or repeated execution. We present ClawProBench, a trace-aware benchmark for runtime-native agent evaluation instantiated on OpenClaw, a live agent runtime with workspace tools and native surfaces for browsing, memory, messaging, scheduling, skills, and subagents. ClawProBench defines two tracks: a 102-scenario full profile with live workspace and native-runtime routing tasks, and a frozen 68-scenario holdout with closed-world JSON output contracts for robust ranking. Trials are scored from execution traces via a safety-gated formula combining correctness, process quality, and efficiency, preserving failure evidence for audit. Our anonymous artifact includes benchmark definitions, scoring code, manifests and sanitized traces. We evaluate 68 configurations on the full profile and 37 on holdout. The top safety-gated average trace score is 0.7671. Native-runtime tasks underperform workspace-live tasks (0.5238 vs. 0.6415). On holdout, pass@k-any outperforms strict three-trial pass (0.6638 vs. 0.2890), while full-profile and holdout rankings show weak alignment (Spearman 0.1300). Rankings based purely on correctness differ substantially from process-aware, safety-gated and strict-pass views. Final-answer leaderboards may hide native-surface weaknesses, one-off successes and trace-local agent failure modes.

Long jumpers with below the knee amputation (BKA) that take off from their running-specific prosthesis (RSP) improved performances significantly over the last years. The long jump biomechanics differs compared to athletes without BKA and the question arises whether the spring-like properties of the RSP facilitate achieving long jumping distances. The aim of this work is to propose a long jump model for athletes with and without BKA, to evaluate it and to apply it for comparing long jump motions with and without RSP. We establish rigid multi-body system models of one athlete with and one athlete without below the knee amputation (BKA). Long jump motions are computed by solving a specific optimal control problem (OCP) with constraints enforcing a physically correct dynamics, both for motion reconstruction or motion synthesis. With the proposed long jump model, we are able to compute realistic long jump motions. We discuss the causes of differences in measured long jumps and show directions for eliminating them. For both athletes, the synthesized solutions reveal potential for performance improvement. The jumping distance of the athlete without BKA is 64cm (6.9%) longer than the one of the athlete with BKA in the synthesized solutions.

Claim-level Uncertainty Quantification (UQ) aims to mitigate the lack of reliability of Large Language Models (LLMs) by evaluating the factuality of each claim in their outputs. We introduce Kernel Token Contradiction (KTC), a lightweight approach to compute claim-level UQ under realistic white-box conditions. KTC represents the candidate tokens involved in LLM generation as a positive semi-definite kernel that integrates both the LLM's conditional distribution and a token contradiction score. We then use the Von Neumann entropy to quantify the uncertainty of this kernel. To estimate token contradiction, we develop a new approach based on frequency statistics from the Wikipedia corpus. Although CPU-only, our approach achieves over an 8.2x speedup compared to state-of-the-art GPU-accelerated methods based on cross-encoders, and over a 65x speedup compared to CPU-only methods with comparable performance. Our evaluation spans two benchmarks across four European languages and 16 different models. KTC not only matches the average performance of existing methods but also outperforms them in high-precision regimes. This combination of computational efficiency and accuracy makes real-time monitoring of LLM outputs practical in production.

Existing AI-for-PDE benchmarks primarily assess models in terms of predictive or approximation accuracy. In physics research, however, AI outputs often serve as evidence for scientific claims. These two objectives are not equivalent: the former measures an output's agreement with a reference target or satisfaction of governing constraints; the latter asks whether, given a specified object of study, scientific claim, assumptions, and evidence standard, the output provides sufficient evidence for that claim. To bridge this gap, we extend a widely used PDE-simulation benchmark and a comprehensive benchmark for PDE inverse problems to enable, for the first time in AI for PDEs, evaluation of whether and to what extent model outputs support specified scientific claims. Our results show that numerical accuracy and evidential support can rank models differently, explain when and why they do so, and reveal that existing benchmarks can favor methods whose outputs provide weaker support for the scientific claims of interest. Together, we formalize, empirically demonstrate, and explain this evaluation–use mismatch in AI for PDEs.

GPU scale-up domains have become the building block of modern machine learning infrastructure, and their design follows a clear trajectory of exponential growth in both interconnect bandwidth and domain size. This paper argues that these two trends are in tension. Through a study of several hundred thousand collective operations across four language models and three recent GPU architectures, we find that GPUs within a scale-up domain arrive at collective barriers hundreds to thousands of microseconds apart, despite executing identical kernels on identical hardware over a uniform fabric. We call this waiting time the synchronization tax and show that it can consume over 50% of collective communication time in an 8-GPU scale-up domain. To understand the sources of this tax, we design a graph-based algorithm that operates on per-rank kernel traces, revealing that cross-rank variation in GEMM kernel execution times accounts for 78% of this overhead. We apply extreme value theory to model this variation and demonstrate that the synchronization tax grows with domain size. Folding this model into an augmented Hockney communication cost model, we show that the synchronization tax fundamentally limits the return on bandwidth scaling and inverts prevailing beliefs about how interconnect bandwidth should scale with domain size.

At the heart of composed visual data retrieval is the fusion of a reference visual input and a textual modification into a single query. While current state-of-the-art methods utilize multimodal large language models for this fusion, their complexity introduces prohibitive querytime latency, limiting their scalability. We instead revisit the efficacy of simple linear interpolation within an embedding space, and introduce SRAIN, the first framework that dynamically predicts query-specific interpolation weights. The key challenge lies in the fact that the quality of an interpolation weight should be measured by the interpolated embedding's discriminability from negatives as well as its proximity to true targets; this makes collecting and predicting optimal weights intractable. We overcome this bottleneck through two key innovations: batch-wise rank-aware weight estimation during training, and a compact memory bank that synthesizes hard negatives during inference. SRAIN achieves the best in composed video retrieval and matches the current state of the art in composed image retrieval, all while substantially reducing querytime latency compared to MLLM-based alternatives.

Baste et al. (2020) conjectured that every regular graph of positive degree has domination number at most its edge domination number, the least size of a maximal matching. Combining published bounds settles the inequality for every degree at least nine. A reduction proves the inequality whenever one endpoint of each edge of a minimum maximal matching can be chosen to form a dominating set, and the Lovász Local Lemma shows such a choice exists for every degree at least seven, newly closing degrees seven and eight and leaving degrees three through six open. The reduction settles each open degree up to a bounded number of vertices, forty-eight for cubic graphs. At fifty vertices, however, the reduction meets an explicit cubic graph it cannot settle, though the inequality holds there too. The inequality cannot be tightened, since infinitely many cubic graphs have equal domination and edge domination numbers. The cubic case stays open, and even linear arguments from the local structure cannot close it. The middle degrees stay open beyond the graphs already settled.

Autonomous underwater vehicles (AUVs) rely on the fusion of inertial navigation systems (INS) and Doppler velocity logs (DVL) for accurate navigation. Before deployment, this fusion requires a DVL initialization pipeline consisting of two stages: alignment, which estimates the rotation between the INS and DVL frames, and calibration, which estimates the DVL error terms. Conventionally, both stages are solved with model-based algorithms that demand complex vehicle maneuvers, surface-level satellite reference measurements, and simplified error models, making initialization time-consuming, trajectory-dependent, and sensitive to sensor quality. In this work, we propose a fully neural- aided DVL initialization pipeline that replaces both stages with two complementary neural networks: ResAlignNet for alignment and DCNet for calibration. The unified pipeline operates in situ on a single nearly constant-velocity trajectory and uses the same inputs as the model-based baseline. Using real-world data recorded across five distinct sensor error-term combinations, the proposed pipeline reduces the velocity root mean squared error by an average of 68.7% over the model-based baseline, using only 25s of data for initialization.

High-altitude platform station (HAPS) serves as a promising enabler for wide-area connectivity of low-power wireless devices, particularly in remote and underserved regions. However, the strong line-of-sight characteristics of HAPS links increase the risk of eavesdropping, while the limited energy budget of ground devices remains a major operational constraint. In this work, we propose a secure wireless information and energy harvesting framework for HAPS-based networks in the presence of spatially distributed eavesdroppers. The proposed system integrates friendly jamming and power transfer nodes equipped with null-steering capability antennas, such that they not only degrade the reception quality at eavesdroppers but also act as additional radio-frequency energy sources for legitimate users. A time-switching wireless information and power transfer architecture is adopted at the user side. Under this framework, we derive tractable expressions for the joint rate-energy coverage and the average secrecy rate using stochastic geometry tools. Numerical and Monte Carlo results validate the developed analysis and reveal key design trade-offs among the time allocation factor, null-steering-zone radius around each user, and jammer transmit power. In particular, the results show that properly coordinated null-steering jamming can simultaneously support secure communication and adequate wireless power transfer, while an appropriate choice of system parameters, such as time allocation factor and jamming power, is required to balance harvested energy, communication reliability, and secrecy performance.

The Jin–Xin relaxation of a nonlinear hyperbolic conservation law introduces a relaxation parameter that controls the width of the internal layer resolving a shock; the discontinuity of the limiting conservation law emerges only in the singular limit as this width vanishes. Physics-informed neural networks (PINNs) use smooth network approximations and are therefore not well suited to this limit, while relaxation PINNs with a fixed parameter resolve only a single scale and cannot follow the multiscale transition toward the limiting solution. We propose the Jin–Xin relaxation gradual convergence method (JXRGCM), which treats the relaxation parameter as a continuation variable, annealing it to zero along a schedule and warm-starting each stage from the previous one, so that the approximation follows the relaxation profile through progressively sharper scales. Under the sub-characteristic condition we establish a stability estimate whose constant is independent of the relaxation parameter; combined with the relaxation limit, it yields for scalar conservation laws an $L^2$ convergence rate of $\mathcal{O}(\varepsilon^{1/4})$ toward the entropy solution. Numerical experiments on the Burgers equation, the shallow-water dam-break problem, and the Sod shock tube show that JXRGCM improves shock and rarefaction resolution compared with fixed-parameter relaxation PINNs and other physics-informed approaches.

Safety alignment helps models adhere to human values, but it often reduces response utility. We ask a critical but understudied question: Does safety alignment impose the cost equally across language groups? To answer this, we introduce a rigorous protocol to measure the utility loss imposed solely by safety alignment, which we term Safety Cost. Through direct pairwise comparisons between safety-aligned models and their unaligned counterparts, we find a systematic inequity: non-English users consistently bear a higher Safety Cost than English users. We further identify three underlying patterns. First, multiple languages lie in a double-penalty zone, experiencing both weaker safety protection and larger utility loss. Second, certain languages exhibit apparent utility gains that are in fact a consequence of safety filters failing to engage. Third, even high-resource languages pay a larger Safety Cost than English to reach the same level of safety. We show that these disparities arise from both explicit refusals and implicit qualitative differences across multiple dimensions. By accurately measuring the disparate effects of safety alignment, our findings expose a systematic disparity in current safety alignment practices.