Papers

Rubric-based evaluation is widely used to assess LLM-based systems by decomposing response quality into task-specific scoring criteria. However, automatically generating rubrics that reliably capture task-specific quality requirements remains challenging. We introduce Mubric, a mutation testing-guided approach to rubric generation. Mutation testing, a classic software testing methodology, evaluates a test suite by injecting faults into programs and checking whether the tests detect them. We draw an analogy between test suites and rubrics: if a rubric captures an important quality requirement, introducing a corresponding defect into an otherwise high-quality response should reduce its score. Mubric first mines common defects from real pairs of preferred and dispreferred responses and abstracts these defects into reusable mutation operators, each specifying how to introduce a particular type of response defect. For a new task, it applies relevant operators to a reference response, checks whether the injected defects reduce response quality, and uses insufficiently penalized defects to refine the rubric. We evaluate Mubric on 703 tasks across four representative domains against six advanced rubric generation methods. Mubric achieves the highest overall evaluation accuracy, outperforming the strongest baseline by 7.48 percentage points.

Power markets are a natural testbed for multi-agent reinforcement learning (MARL), where multiple self-interested participants repeatedly submit bids. A market-clearing mechanism then determines dispatch and prices subject to power grid constraints and market settlement rules. However, existing MARL environments typically focus on a single market setting, implement simplified clearing mechanisms, or rely on CPU-based optimization solvers that slow large-scale training and limit the systematic study of bidding strategies and market behavior. We introduce PowerMarketJax, a benchmark suite for MARL across five power markets: day-ahead wholesale, real-time balancing, ancillary services, peer-to-peer double auctions, and local flexibility. Each environment implements its own clearing, pricing, and settlement rules while providing a common framework for learning and evaluation. We find that learned bidding behavior depends strongly on the market design: independent learners can miss better strategies when gains require many agents to change together, when more profitable strategies lie beyond a region of lower profit, or when profits disappear as more agents adopt the same strategy. PowerMarketJax implements both market simulation and policy training in JAX, allowing the entire pipeline to run on the GPU with 1,024 X 1,200 parallelisms across both environments and market participants, achieving up to 33X speedup over CPU-based baselines. Our open-source benchmark is available at: https://github.com/powermarketjax/PowerMarketJax.

The rational identity testing problem, RIT, asks whether an input rational circuit, computes the zero element of the free skew field. For rational formulas the problem is known to be in deterministic polynomial time. For rational circuits of size $s$ the associated linear pencil has dimension $2^{O(s)}$, and the known algorithms use exponential time and exponential space. We show that RIT for rational circuits over $Q$ and finite fields is in PSPACE. As a consequence, noncommutative PIT for degree unrestricted noncommutative circuits is also in PSPACE. The proof is based on the following three observations Analyzing the Hrubes-Wigderson construction of a linear pencil for an input rational formula, we obtain a succinctly represented linear pencil $A$ of size $2^{O(s)}$ for the input rational circuit of size $s$. More precisely, given indices $i$ and $j$ for the linear pencil we can compute $A_{i,j}$ in space polynomial in $s$ and length of $i,j$. This algorithm essentially gives a succinct presentation of the linear pencil $A$ obtained by the reduction. By the recent theorem of Chatterjee, Ghosh, Gurjar, Raj and Thierauf that deciding whether a symbolic matrix $\sum_i A_ix_i$ has full noncommutative rank is in NC. We note that their algorithm is logspace-uniform NC and we use it as a black-box for computing rank of succinctly represented pencil $A$. Finally, we note a folklore simulation: If a problem is solved by a logspace-uniform family of deterministic Boolean circuits of polylogarithmic depth, and its input is not written down but is presented by a polylogspace subroutine that returns any requested input bit, then the circuit can be evaluated in polylogarithmic space. Hence, simulating a depth $O(\log^{i}M)$ circuit for succinctly presented inputs of length $M = 2^{\Theta(s)}$ yields a polynomial space algorithm.

Omnimodal evaluation should go beyond independent text, image, and speech production: individually plausible outputs may not express a coherent shared event. We introduce Omni-StoryBench, a story-grounded omnimodal benchmark evaluating whether models can coherently continue stories across image, narration, and speech. Each instance provides a current storybook page and structured next-page conditions, requiring models to generate the next illustration, narration, and spoken character utterance. Omni-StoryBench contains 900 rigorously validated story transitions from openly licensed children's books, with ground-truth next-page references and speech metadata. We evaluate systems with modality-specific metrics and consistency-centered LLM-as-a-judge rubrics for context preservation, condition following, reference consistency, and cross-modal coherence. Across 32 baseline configurations spanning orchestration, semi-orchestration, and native any-to-any paradigms, we find orchestration with strong VLM planning most reliable, while current native omnimodal models often struggle with output completeness and controllability. Our analysis shows text-side performance is associated with image and speech quality, but image generation and visual continuity form the clearest observed bottleneck among the evaluated configurations. These results position Omni-StoryBench as a system-level benchmark measuring coherent omnimodal generation beyond isolated modality quality.

Agile flight tasks such as drone racing and pursuit-evasion require strong acceleration and precise turns, but the available thrust changes as the battery discharges and voltage drops under load. Conservative command limits make this variation easier to tolerate, at the cost of unused performance. We investigate how learned controllers can use that additional thrust while retaining the flight controller's voltage compensation and rate control. Our training simulator couples an identified load-transient battery model to rotor dynamics and firmware saturation. The feedforward policy receives filtered voltage during both training and deployment. Controlled ablations distinguish the benefit of a larger thrust-command range from that of voltage information. On a 38 g Crazyflie Brushless, the resulting policy reduces circle tracking error by 49% relative to stock-authority RL at 3.84 m/s, while preserving easy-task precision. Mean 20-lap race time decreases from 106.22 s to 95.24 s. Compared with a voltage-blind policy with the same increased authority, hardware error and race time are lower by 15.3% and 4.5%, respectively. In simulation, replacing the policy's voltage input with a recording from a different battery condition worsens hard-circle tracking, with a smaller, voltage-dependent effect in racing. Together, these results show where a simple voltage input complements existing actuator compensation in aggressive learned flight.

Tool-using agents are entering settings where a wrong action carries real cost, and the benchmarks certifying them grade what each simulated tool call reports having done, assuming the tool did what its interface advertises. The audit taxonomies we survey publish no category for that assumption, and a defect beneath a score is present on every rerun. We treat a tool's advertised surfaces as an executable contract, check the implementation against it, and trace each score's provenance through the task files and evaluator code to the verdicts that derive from state a defective tool should have written. Across 34 audited mutating tools in four benchmarks we confirm seven tool defects and one evaluator property at pinned commits. On injected defects the checker raised no false positive in 25 flags, flagged 2 of 5 negative controls, and missed most: in 29 of 33 scored misses a clause covered the defect but no probe revealed it. The checker's own static half, run alone, flags 14 of 17 confirmed sites, so on these findings the dynamic half confirms and traces rather than discovers. Twelve further AgentDojo tools, with six held-out tools and the seven audited first, complete its 25-tool mutating surface, on which at least 5 tools diverge from their advertised surface as our contracts read it, a rate for AgentDojo alone. No gold trajectory reaches either tau2-bench defect; on 1,120 paths built to isolate the telecom defect, a number fixed by construction, the evaluator rewards a refuel of a suspended line and fails the repaired tool. The clearest case is a clinical benchmark whose tool tells the agent each write executed under a documented no-write design its interface does not disclose; its grader takes that message as evidence, so its action success rate records whether a request carried the expected payload, not whether any record changed.

Synthetic anomaly generation helps expand industrial anomaly datasets when real defects are scarce or unavailable. Existing approaches lie at two extremes: procedural approaches are fast but struggle to represent complex anomalies, while generative approaches produce diverse defects but require costly per-sample generation. We present FLASH, a framework that decouples defect generation from anomaly synthesis under a “generate once, synthesize many” paradigm. Given only normal images, FLASH uses Vision-Language Model (VLM) guidance and an image-generation model to produce a small set of defect images, from which it extracts, validates, and banks reusable defect patches. For synthesis of anomalous images, Object Boundary Suppression (OBS) first identifies the probable foreground object-aware region of the host image, while Multi-Resolution Spectral Pyramid (MRSP) noise generates diverse, size-controllable masks that determine the defect location and spatial extent. It then composes a large and diverse synthetic anomalous image set by localizing the defect region, sampling size-controllable placement masks and seamlessly blending retrieved defects onto new defect-free images without further need for image generation. Experiments on the MVTec AD 2 dataset show that FLASH-generated anomalies nearly close the calibration gap on real defects, reaching 78.1% image-level F1 against an 83.6% real-anomaly upper bound and providing the most consistent calibration transfer across detectors among procedural and generative alternatives. Moreover, FLASH synthesizes anomalies more than 11.95x faster than per-sample generative approaches.

Can reasoning models trick chain of thought (CoT) monitors and perform hidden computation without revealing it in their thinking traces? We show that the answer depends on the underlying task difficulty and the model size. Simple computations can be performed covertly; however, beyond a threshold depending on model size, successfully solving the task necessarily leaks a near-linear amount of information about the covert task input into the CoT. Therefore, sufficiently complex hidden computation always leaves an information-theoretic footprint. However, concerningly, this leakage need not be readable: Under plausible cryptographic assumptions, even a one-layer Transformer can encrypt its reasoning online so that no polynomial-time monitor can extract information about the hidden computation. Overall, our theoretical and empirical results provide a holistic view of both the opportunities and the limitations of CoT monitoring.

Recent Recommendation Agents (RecAgents) offer a promising alternative by shifting recommendation to an active, user-side paradigm, where generative agents autonomously perceive external platforms, reason over user preferences, and execute decisions. However, existing RecAgents still suffer from two critical limitations: brittle item perception based on noisy and heterogeneous item pages, and inefficient long-context reasoning over extended user histories and multi-step interaction traces. To address these challenges, we propose a novel recommendation agent framework, termed as ReMem, that combines OCR-based multimodal perception with time-evolving dynamic memory. Instead of parsing raw HTML, ReMem observes item pages through screenshots and extracts structured multimodal information via an OCR tool, enabling a more humanoid and platform-agnostic perception mechanism. To support long-horizon preference modeling, ReMem further introduces a chunk-wise sequential memory update strategy, where the agent selectively maintains a fixed-size memory of informative historical interactions while processing arbitrarily long contexts with linear inference complexity and bounded context length. This design allows the agent to preserve evolving user preferences without relying on external memory modules or disrupting the standard autoregressive generation process. To enhance the dynamic memory instruction, we further develop a multi-memory GRPO variant, which propagates the final-answer advantage to all intermediate conversations that contribute to the final response. Extensive experiments on three datasets demonstrate that ReMem consistently outperforms state-of-the-art baselines, achieving an average improvement of 5.16% across three recommendation agent tasks, namely searching, ranking, and judging.

With LLM watermarking being deployed commercially and now required by regulations, improving its reliability and effectiveness has become crucial. Yet, recent progress in the field of LLM watermarking has increasingly been driven by improving details of existing methods, an effort fundamentally limited by the pace of human researchers. In this work, we enable for the first time the autonomous discovery of new distortion-free state-of-the-art watermarking schemes. To enable this, we (i) establish strict criteria to ensure that watermarks are reliable (e.g., they do not have an unexpectedly high false positive rate), (ii) propose rigorous statistical tests to automatically evaluate whether a watermarking scheme satisfies our criteria, and (iii) design an evaluation suite to rank watermarks along three key dimensions: detectability, quality, and robustness. By running our framework with 3 frontier models (GPT-6 Astra, Opus 5, Gemini-3.8 Flash), we discover over 50 different watermarking schemes, including several that outperform prior works along all key dimensions. We complement this by a manual study of the discovered schemes, distilling the key ideas into smaller components, and individually studying the impact of each component across dimensions (detectability, quality, robustness) to better understand how the proposed schemes operate. Importantly, we find that the agents, on top of improving existing ideas, also discover fundamentally new ideas (e.g., aligning watermark scores with random per-request direction). Overall, our work establishes the first steps of fully autonomous watermarking research, enabling the discovery of more reliable and effective watermarks. Our code is available at https://github.com/eth-sri/automark, and a blogpost to visualize our results at https://www.sri.inf.ethz.ch/blog/automark.

This paper introduces a hybrid joint-selective optimization (HJSO) framework for large-scale numerical problems in which a small subset of trainable quantities is of primary interest. We partition the full parameter vector into a high-dimensional remaining block and a low-dimensional block of parameters of interest (POIs), perform joint first-order optimization over the full parameter set, and then freeze the remaining variables while applying a reduced-space Levenberg-Marquardt (LM) refinement to the POIs. The method is designed for settings in which the POIs are low-dimensional but strongly influence the quality of the computed solution, while the full parameter space remains too large for full-space second-order methods. The framework is evaluated on three representative problems: a matrix eigenvalue problem, an inverse Bratu problem solved with a physics-informed neural network, and a 100-dimensional nonlinear Black-Scholes problem solved with the DeepBSDE method. In each test, HJSO reaches prescribed POI-error thresholds faster than the corresponding joint first-order baseline and improves the final POI accuracy for the reported solver configurations. The contribution is therefore not a universal optimizer, but a practical reduced-space strategy for problems with known low-dimensional parameters of interest and expensive high-dimensional training variables.

Vision-language-action (VLA) models have driven rapid progress in robotic manipulation, demonstrating strong fine-grained control and promising performance on long-horizon tasks. However, many existing VLAs lack explicit access to interaction history, making them vulnerable to perceptual aliasing: similar current observations and robot states at different task stages may induce action ambiguity and lower success rate. Existing methods incorporate temporal or progress cues through feature conditioning, action-prior modification, or sampling guidance. However, methods that jointly fine-tune memory modules and the base VLA incur additional policy-training costs, motivating the separation of trainable history-conditioned steering from frozen base-policy refinement. We propose ActMem-VLA, a dual-expert handover architecture that augments a frozen, fine-tuned VLA with a memory plugin comprising a Mamba-based memory module and a lightweight PreAction Expert (PAE). Specifically, Mamba encodes executed-action history into memory that conditions PAE alongside current context. With these inputs, PAE steers task progression during early, high-noise denoising, then passes the partially denoised action to the frozen Action Expert (AE) to refine action details during the remaining low-noise steps. The fine-tuned base VLA remains frozen throughout training, while only the Mamba module and PAE are jointly optimized. On LIBERO-Mem, ActMem-VLA achieves 80.8% average success across all ten tasks, compared with 65.2% for $\pi_{0.5}$ and 49.5% for MemoryVLA, while introducing only 3.45% additional parameters. Across four real-world tasks, it improves the average success rate over $\pi_{0.5}$ by 28.8%.

Large reasoning models improve performance on challenging problems by allocating additional computation before answering, but longer reasoning does not always lead to better results and can introduce substantial redundant reasoning on simple problems. Conversely, aggressively shortening reasoning can degrade performance on difficult ones. Effective reasoning therefore requires dynamically deciding when additional computation is useful based on the reasoner's capabilities and evolving solution state. Existing approaches often rely on predefined budgets or intervention rules, retrain the target reasoner, or require additional supervision. We introduce MetaCtrl, a lightweight controller that adaptively regulates a frozen reasoner without predefined token budgets or reasoner retraining. We formulate reasoning regulation as a sequential metacognitive control problem: MetaCtrl observes the evolving reasoning trace and decides whether to continue, simplify, skip redundant steps, or conclude reasoning. It is trained directly with reinforcement learning using a reward that prioritizes correctness while favoring shorter trajectories among correct solutions, requiring neither supervised intervention trajectories nor problem-specific budgets. Across seven benchmarks spanning mathematics, science, and code, MetaCtrl consistently improves the accuracy of LRMs while reducing their reasoning length. On DeepSeek-R1-Distill-Qwen-7B, it improves average accuracy by 4.7 points while reducing generation length by 53.3%. Without further training, the same controller transfers to an unseen reasoner (e.g., Qwen3-14B), improving average accuracy by 2.9 points and reducing generation length by 50.3%. These results establish MetaCtrl as a plug-and-play controller for improving reasoning accuracy while substantially reducing inference-time generation. The code is available at https://github.com/binbin2xs/MetaCtrl.

Deep vision models often degrade under distribution shift. Test-time adaptation can improve robustness but typically requires iterative optimization, hyperparameter tuning, and multiple forward-backward passes. We propose Zero-Training Fisher Geometry Alignment (ZFGA), a closed-form method that improves robustness under covariate shift without modifying model parameters. ZFGA is based on the observation that distribution shifts distort feature-space geometry. It estimates the Fisher information matrix of the predictive distribution with respect to feature embeddings and applies a linear transformation that aligns test-feature Fisher geometry with a reference geometry computed from clean data. This provides a natural-gradient-inspired preconditioning step in feature space. We evaluate ZFGA on CIFAR-10-C and ImageNet-C using ResNet-50, DINO ViT-S/16, and CLIP ViT-B/32. ZFGA consistently improves over zero-shot inference across all three models, although it is not the strongest method for every model. Covariance whitening performs better on ResNet-50, while Fisher whitening is statistically indistinguishable from ZFGA on CLIP. Across six training-free and gradient-based alternatives (covariance whitening, Fisher whitening, TENT, T3A, LAME, and AdaNPC), ZFGA is the only method that does not substantially harm any of the three model families. The Fisher geometry distortion is also positively correlated with ZFGA gain (Pearson r = 0.366, p = 0.017), providing preliminary evidence that geometric misalignment contributes to robustness degradation. ZFGA requires only forward passes and matrix operations at inference time, offering a lightweight and deterministic alternative to optimization-based test-time adaptation.

Conformal prediction provides set-valued predictions with distribution-free coverage guarantees, making it attractive for high-stakes image classification. However, split conformal prediction is data-inefficient, while full conformal prediction (FCP), despite its stronger statistical efficiency, is computationally prohibitive at scale because it requires candidate-specific model refits at test time. We address this limitation by leveraging zero-shot vision-language models (VLMs) to guide scalable FCP in large label spaces. We introduce Targeted Full Conformal Prediction (T-FCP), which uses a lightweight inductive conformal predictor to prune unlikely labels and applies FCP only to the remaining candidates, reducing computation while retaining the formal guarantee of the combined conformal procedure. We further propose Stabilized Online LDA (SO-LDA), an efficient VLM adaptation solver based on rank-one inverse-covariance updates. Across multiple benchmarks, including ImageNet, T-FCP enables practical full-conformal image classification with modest test-time overhead, yielding efficient prediction sets and more stable empirical coverage than split conformal alternatives.

Cross-skeleton motion generation trains generative models to carry action structure and motion intention from one body to another. Yet a target motion that shows the right action has two explanations that the training data cannot tell apart: the model transferred the source clip, or it recovered a typical motion for the requested action. We show that this ambiguity is structural rather than incidental: under standard generative objectives, the source-conditioned retargeting map is non-identifiable in sparse heterogeneous motion domains. Unpaired distribution matching yields gauge non-identifiability: the latent spaces of different skeletons can be transformed relative to one another without changing the training evidence, so different source-conditioned maps fit it equally well. Sparse paired supervision admits the complementary failure mode, conditional-mean degeneration: when clips are paired only by action, squared-error training converges to an average target motion that ignores the source clip. To make the missing evidence observable, we introduce Source-Instance Fidelity (SIF), a diagnostic that tests whether outputs differ from one another the way their source clips do, with the target skeleton and action held fixed. Under this diagnostic, methods that succeed at the standard action-level test on animal motion data often sit at the source-blind floor, while the methods that rise above it retain only a partial relational signal. Retargeting therefore needs objectives and evaluations that can identify the source-conditioned map it claims to learn. Project page: https://cross-skeleton-retargeting.netlify.app/.

Users are central to transportation systems, yet their roles are often simplified in transportation modeling and decision-making. Conventional approaches primarily represent users through demand-related inputs, such as trip flows, delivery requests, and charging loads. However, digital, electrified, and platform-based services create more direct user-system interactions, with users responding to prices, incentives, service availability, and information in ways that can influence operational and planning decisions. This review develops a concept-oriented, taxonomy-based synthesis of how user-related information is represented and used in transportation decision-making. Rather than organizing the literature primarily by transport domain, we distinguish three user roles: passive demand, operational agents, and system co-shapers. Drawing on studies across transportation electrification, urban logistics, shared mobility, multimodal transport, and Mobility-as-a-Service (MaaS), the review examines how user preferences, flexibility, willingness to pay, acceptance, participation, and feedback are incorporated into modeling and decision frameworks. The synthesis shows that progress has been made in measuring and predicting user behavior, but user-related variables are still often used mainly for behavioral analysis rather than as actionable inputs to system decisions. Their integration into operational decision-making remains incomplete, and systematic feedback from operational user responses and outcomes to long-term planning is even more limited. Key challenges include closing the loop between operation and planning, modeling heterogeneous and strategic users, integrating optimization with user acceptance, and ensuring privacy, equity, and deployability. Together, these findings point toward a shift from user-behavior analysis to user-informed and user-driven transportation system design.

Large language models (LLMs) are increasingly used for software engineering tasks, yet their stochastic behavior challenges the validity, reproducibility, and comparability of their evaluations. Conventional practices such as reporting a single output, an average score, best-of-N, or pass@k performance can obscure variability and estimation uncertainty, potentially leading to misleading conclusions about system reliability. This article presents SafeLLM4SE, a practical methodology and reporting standard for statistically principled evaluation of LLM-based software engineering systems. Rather than treating generated outputs as deterministic artifacts, SafeLLM4SE treats them as realizations of a stochastic process and distinguishes quality, stability, and estimation uncertainty. It combines adaptive sampling with confidence intervals, distribution-aware statistical comparisons, effect sizes, and a minimum reporting standard covering model configuration, reproducibility, evaluation procedures, and resource usage. SafeLLM4SE is also provided as an open-source software package available on PyPI, enabling researchers and practitioners to reproduce and extend the methodology. We illustrate its application by comparing two LLMs on HumanEval, a benchmark of programming problems assessed through functional tests.

Physical active vision allows robots to change their viewpoint when task-relevant observations become unreliable, yet existing manipulation benchmarks provide limited support for studying how policies recover from occlusion during execution. We introduce BAVO-Bench (Bimanual Active Vision under Occlusion), a bimanual active-vision benchmark that systematically controls external visibility through Clean, Stage Occlusion, and Random-time Occlusion conditions, enabling evaluation of both manipulation performance and active visual recovery. Building on this setting, we present A-FAR (Active Future-Aware Recovery), an active-vision policy for joint viewpoint and manipulation control. A-FAR represents moving-camera observations in a unified robot-centric 3D frame and distills relational structure together with its future evolution from a pretrained 4D model, providing the policy with future-aware geometric guidance without requiring future observations at deployment. Experiments across multiple manipulation tasks show that A-FAR improves robustness to both structured and temporally shifted occlusions while maintaining strong performance under clean observations.

This paper presents a comprehensive methodology for vertiport design and capacity analysis using stochastic queueing models. Vertiport capacity is defined as the maximum steady-state throughput of aircraft per hour while maintaining a specified maximum queuing delay with high probability. We introduce a two-node queueing network model with a finite buffer to analytically derive mean queue lengths and delays at singleTLOF vertiports, supported by a simulation model that validates these findings under various operational conditions. Our analyses reveal that vertiport capacity is significantly constrained by the interactions between takeoff/landing operations and turnaround activities. Capacity is further reduced by the variance in these operational parameters, with deterministic times substantially enhancing capacity. The study shows that vertiport throughput is typically around 50% of the theoretical maximum departure capacity in a Markovian model, decreasing sharply as confidence in maximum queuing delay increases. Non-Markovian models with adequate parking pads achieve full TLOF capacity. Additionally, our methodology identifies the optimal number of parking pads required to maximize throughput per unit area. The findings provide insights for designing vertiport infrastructure to support scalable and efficient UAM deployment.

Image translation is a fundamental capability of multimodal models for multilingual applications, requiring visual understanding and meaning preservation across languages. However, existing benchmarks have limited language coverage and often lack explicit image-specific evaluation criteria, making it difficult to comprehensively assess this capability. To systematically evaluate this capability, we introduce VISTA-Bench, covering 22 languages and 10 domains, and develop an image-specific rubric evaluation protocol. The benchmark combines sampling for language and scenario coverage with model-assisted, human-verified annotations that group related text into coherent semantic units and provide multilingual reference translations. The rubrics specify essential content, semantic relations, and acceptable translation variants, yielding separate output-based scores for translation quality and the preservation of visual and knowledge-dependent information. We conduct extensive evaluations of 16 mainstream models, including 12 multimodal models and four text-input models, and provide systematic analyses across languages, domains, and evaluation dimensions.

Laboratories often face a new molecular assay with 16-64 labels and a bank of predictors whose training data and parameters are unavailable. The practical question is which frozen outputs to include in a small local model. AssayRouter treats completed assays as pseudo-targets and labels each candidate by its post-fit utility: the reduction in held-out discovery loss when the candidate is added to the local target predictor. A shared regressor learns to predict this utility from candidate behavior on the support set, without source identity; on a new assay, one frozen ranking selects four sources and separate labels fit a convex combiner. We train only on completed ChEMBL-MT assays and evaluate 24 external regression assays across six frozen interface families. AssayRouter-C lowers strict four-call negative log-likelihood (NLL) by 0.0409 relative to Support-CV@4. Frozen candidate-label permutations confirm that candidate-utility correspondence carries the transferred information, and leave-one-interface-out training shows that the mapping generalizes to unseen predictor families. Completed assays therefore provide transferable supervision for scarce-label routing through frozen prediction interfaces.

Medical multimodal large language models (MLLMs) are increasingly expected not only to answer clinical questions, but also to localize the visual evidence behind their predictions. A common strategy connects a vision–language model (VLM) with SAM-style segmentation through a special <SEG> token, yet full-parameter training of this unified architecture is difficult because image-level reasoning and pixel-level segmentation impose different requirements on the shared representation space. To address this issue, we propose a parameter-decoupled training framework for unified medical reasoning and segmentation. The framework treats the <SEG> hidden state as a semantic-to-spatial prompt for the mask decoder and encourages it to become separable from generic language states, reducing ambiguous segmentation prompts and potential disruption to reasoning representations. It first performs medical shallow alignment to adapt visual features to clinical language without disturbing the LLM; then controlled instruction tuning shapes separable <SEG> prompt states, monitored by the Davies–Bouldin Index (DBI), while scaling segmentation gradients entering the language backbone; finally, the SAM branch is specialized with the VLM frozen to improve mask precision without altering reasoning parameters. Experiments on medical referring segmentation, grounding, visual QA, and textual QA benchmarks show that our framework achieves strong language-conditioned segmentation while preserving competitive reasoning ability. Ablations show that two-phase instruction tuning, gradient scaling, and segmentation specialization all contribute to the model.

AI-assisted cheating has become an important threat to the security of online exams. This study examines whether the risk of AI-assisted cheating in the final exam can be predicted using students' digital traces in the learning management system (LMS) during the first eight weeks of the semester. The sample comprised 52 first-year undergraduates enrolled in a bachelor's program in Computer Education and Instructional Technology and taking an Introduction to Programming course at a public university in Turkiye. Students were labeled as low- or high-risk based on suspicious behaviors recorded in the final-exam logs, including copy, focus-loss, and right-click events. Of the 52 students, 23 (44.2%) were labeled as high-risk in a proctored, face-to-face exam. Group membership was then predicted using five features selected from 27 candidates extracted from students' digital traces. Logistic Regression, Naive Bayes, Random Forest, and Gradient Boosting algorithms were used to build the prediction models. Model performance was evaluated using leave-one-out cross-validation (LOOCV) with fold-specific preprocessing and feature selection. Logistic Regression achieved the best performance (Accuracy = 73.1%). The results indicate that LMS interaction data can provide an early signal of AI-assisted cheating risk. Course-module views, assignment submissions, and the number of days on which course videos were accessed were the most consistently selected features across the LOOCV folds. These predictions are intended to support timely academic guidance, not to establish misconduct or initiate disciplinary action.

Neural solvers offer efficient surrogates for numerical simulation of partial differential equations (PDEs). For time-dependent problems, strong one-step accuracy does not necessarily translate into reliable autoregressive rollout. We observe that a solver based only on physical-state modeling can achieve lower one-step error, whereas its spectral-only counterpart can become more accurate at later rollout steps. Motivated by this observation, we present Transolver-$\sigma$, a neural PDE solver based on joint spectral–physical subspace modeling. Within each block, adaptive physical-state interactions and spectral transformations are modeled in dedicated latent subspaces, whose responses are recomposed to enable information exchange between the two representations. Within the physical subspace, we introduce Slice-Residual Physics-Attention (SRPA), which preserves an explicit slice-space identity path while retaining learnable cross-slice interaction. In parallel, an axis-factorized Fourier operator captures global spectral structure. Across five well-established PDE benchmarks spanning steady-state prediction and time-dependent dynamics, Transolver-$\sigma$ achieves state-of-the-art with a benchmark-averaged relative error reduction of 33.4% over the strongest baseline for each metric, while consistently improving autoregressive rollout over single-operator counterparts. Transolver-$\sigma$ further delivers strong gains on coupled multiphysics systems and real-world fluid and combustion measurements from RealPDEBench, demonstrating its effectiveness beyond standard simulation benchmarks.

In this paper, we study the matroid base packing problem from the perspective of cooperative game theory and define the associated matroid base packing game. This model extends the network strength game of Baiou and Barahona (2020) from graphic matroids to general matroids. Building on the exchange structure of matroid, we develop a graph theoretic framework for analyzing the core and the nucleolus in this game. This framework yields a compact characterization of core nonemptiness and enables polynomial time algorithms for testing core nonemptiness, finding a core allocation, and computing the nucleolus. In particular, the approach provides a broader and more structured framework for the nucleolus computation problem arising in network strength games studied by Baiou and Barahona (2020), covering both the nonempty core and empty core cases.

The acquisition of high-dimensional channel state information (CSI) in wireless MIMO-OFDM communications usually requires high pilot overhead, or relies on accurate and complete positional or environmental information. In this paper, we propose a geometry-aided channel deduction (GCD) approach, which utilizes an uncalibrated digital twin (DT) with only approximate environmental geometry and positions to assist the channel acquisition. The key rationale behind is that, even imprecise geometric information, which can be easily obtained in advance through radio sensing technologies or existing geographic databases, provides certain structural features about the current channel; meanwhile, the coarse instantaneous channel estimates using only a small amount of pilots provide dedicated information that aligns with the channel structure and further compensates for the geometry inaccuracy and other channel unknowns. To this end, we first extract geometric features from the DT, which contain only simple structural information of the channel. Then we propose random prompt augmentation, a novel method to generate an appropriate prompt that converts geometric multi-path structure into a CSI-like representation while suppressing the disturbance of other unknown channel parameters. The prompt is then fused with the pilot-based instantaneous channel estimate via a channel deduction network. To further enhance the network's versatility, we incorporate pilot configurations into the existing learning architecture to support variable pilot patterns. Comprehensive experiments validate the superiority of the proposed method, which demonstrates high channel acquisition quality, low pilot overhead, and strong robustness. Furthermore, the structural prompt also serves as scenario-related context, enabling our approach to generalize well in new scenarios.

We initiate the study of quantum leakage resilience of unmodified Shamir secret sharing over prime fields. A well-studied leakage model for Shamir's secret sharing classically is single-bit local leakage from each share. We consider its quantum analogue where, for each party, a local leakage channel takes as input the party's share and outputs a leaked qubit. Without preshared entanglement, we show that the distinguishing advantage is $2^{-\Omega(n)}$ when the threshold rate $t/n=\tau$ exceeds $\tau_\star\approx0.73339$ by a fixed positive margin. More generally, we allow disjoint entangled blocks of any fixed maximum size where there is no entanglement between different blocks or with the adversary, and each block emits at most a fixed number of qubits. Security holds when the threshold rate is high enough (sufficiently close to one). We then allow a specified set of devices to share entanglement with the adversary. We show that security holds even when a linear number of devices ($\alpha n$ for small $\alpha>0$) share entanglement with each other and with the adversary for a large enough threshold rate. As a complementary negative result, we also show that even classical single-bit leakage makes Shamir scheme insecure if we allow arbitrarily large entanglement between the leakage devices. A GHZ state shared by exactly $t$ leakage devices makes even classical one-bit leakage insecure, without any entanglement with the adversary. In this attack, each participating device emits only one classical bit, and their joint parity distinguishes any chosen pair of secrets with a constant advantage. Thus, for fixed threshold rates above $\tau_\star$, the maximum number of devices that may share arbitrary entanglement with one another and with the adversary while preserving security is linear in $n$ up to constant factors, although the optimal support fraction remains open.

Non-injective mappings in neural networks map distinct inputs to the same representation, thereby implicitly inducing equivalence relations in the input space. However, the input differences eliminated by these mappings may still be required by downstream tasks, creating a mismatch between operator-induced indistinguishability and task-required distinctions. For non-injective linear operators realized in the current forward pass, their null spaces exactly characterize these invisible input variations. We propose Task-Relevant Null-Space Residuals (NSR), a general residual framework for non-injective linear mappings. NSR combines null-space component extraction from pre-mapping representations, member-level encoding and gating, and application-specific integration to exploit potentially task-relevant information under downstream supervision while preserving the original aggregation or merging rules. We evaluate NSR in two structurally different settings: token merging and graph aggregation. In token merging, NSR achieves higher semantic segmentation performance than the corresponding compressed baselines in 34 out of 36 evaluated configurations, with a maximum observed gain of 31.51 mIoU points under strong compression. In graph aggregation, NSR achieves 100% training accuracy on Tree-NeighborsMatch at depths d=2–6 across three backbones, alongside gains on heterophilic node classification and molecular graph regression. Together, these results support null-space residuals as a practical complement to non-injective linear mappings, enabling downstream models to learn from input distinctions invisible in the original operator's output.

Proactive LLM agents can turn idle compute into useful support before users ask. Yet even correct work can misread user context, impose review costs, or undermine trust. This work proposes foundations for designing, realizing, and evaluating proactive LLM agents around three joint principles (3T): Task Capability, anticipating relevant needs and correctly performing useful work; Temporal Allocation, allocating compute according to resource availability and when results are needed; and Trust, sustaining users' confidence and appropriate reliance on the agent. We connect these objectives to a design space organized around five dimensions: task scope, anticipation horizon, activation trigger, processing timing, and intervention depth, and specify the situation and system modeling needed to support its choices, including user and environment representations, backbone LLMs, and agent harnesses. Lastly, we propose PROACTIVITY-GYM, a simulation-based evaluation testbed including multi-day scenarios, stateful environments, and persona-conditioned simulated users that can evaluate the consequences of proactive assistance across interactions. Evaluations across 23 model-harness configurations uncover substantial performance gaps across 3T and reveal that LLM judges often conflate task capability and trust. A human study with 30 participants demonstrates the importance of the joint 3T optimization: participants show sharp trust declines after intervention misalignment despite correct outcomes, and prefer sleep-time assistance, even when imperfect, to preserve ongoing focus. Together, these findings support designing and evaluating proactive agents through the joint consideration of useful work, compute allocation, and evolving user trust.