arXiv:2607.00420v1 Announce Type: new Abstract: Organizations increasingly deploy autonomous artificial intelligence (AI) systems for operational decisions, such as inventory replenishment. Yet fully granting override rights can degrade performance due to human bias and noise, while prohibiting them may overlook valuable private information. This raises a key question: How should override rights be structured to improve human supervision of autonomous AI? Methodology/results: We propose a constrained override policy that limits overrides per decision episode to enable selective filtering that prioritizes high-value overrides. We tested it through a randomized field experiment with 553 workers at a major Chinese smart vending machine retailer that manages more than 59,000 machines and 4,000 SKUs. Workers were assigned to no overrides, free overrides, or a two-per-machine limit on downward overrides. Free overrides reduce inventory by 1.95% but also cut sales by 1.19%. Constrained overrides reduce inventory by 1.28% without harming sales, as workers select better SKUs to override, confirmed via local average treatment effects. Gains are largest for experienced workers, high-incentive SKUs, and growth-stage SKUs. A simulated personalized policy further increases sales probability by 9.1%. Managerial implications: Academics gain novel insights from the causal effects of discretion design in human-supervised AI, emphasizing selective filtering to enhance decision quality. Managers can benefit from a scalable, low-cost policy for operations such as retail, logistics, and resource planning, reducing excess inventory without sales loss while harnessing private human information, with no need for algorithmic redesign, information customization, or additional training.
Science Journals
arXiv:2607.00139v1 Announce Type: new Abstract: The cost of human expert evaluation is a principal bottleneck to deploying language models in specialized, high-stakes domains. This is particularly acute for Arabic sociolinguistic knowledge: credible grading requires not only linguistic fluency but deep cultural familiarity that cannot be approximated by surface-level metrics. We address this with a cross-evaluation framework instantiated on two underrepresented Arabic dialect communities: Egyptian and Iraqi Arabic. We contribute 103 validated prompt-rubric pairs (70 Egyptian, 33 Iraqi; 53 Cultural, 50 Linguistic), authored and graded by native-speaker SMEs using penalty-weighted rubrics distinguishing positive content requirements from answer-specific negative error criteria. Three frontier LLMs serve as target models (graded by human SMEs across 302 unique prompt-response pairs), while five frontier LLMs serve as automated judges enforcing a provider-level self-evaluation guard. A dual-metric scheme combining Mean Absolute Deviation (MAD) with Signed Mean Error separates directional grading bias from symmetric noise. Across 1,307 judge evaluations: GPT-5.4 is the most reliable judge (MADj = 10.21 pp, Signed Error = -1.12%); four of five judges show systematic leniency (+2.01% to +6.56%); Cultural tasks are harder to grade than Linguistic tasks for all judges (MAD gap 1.83-4.78 pp); and models substantially outperform on Egyptian prompts compared to Iraqi prompts. However, given leniency differences between Iraqi and Egyptian SMEs, we cannot solely attribute this gap to model knowledge. We therefore emphasize findings that do not assume identical leniency across human graders. Across all samples, implicit cultural reasoning -- requiring models to simulate native-speaker judgment rather than rely on lexical verification -- emerges as the primary failure mode for automated grading across all judge models.
arXiv:2607.00808v1 Announce Type: new Abstract: Pre-training on large-scale videos to improve reinforcement learning efficiency is promising yet remains challenging. Existing methods typically treat the agent as an indivisible entity, modeling motion patterns globally. Such global modeling is tightly coupled with the morphology, hindering transfer across domains. In contrast, despite the vast disparity in global motions, the local components exhibit similar motion patterns across different agents. Building on this insight, we propose a novel Deconstruct-Recompose Paradigm (DRP) for learning transferable local motion representations. Specifically, in the Deconstruct phase, we identify multiple local points and track their frame-wise motions, defining each as an Atomic Action. We introduce a Dual-Attention Encoder (DAE) to learn local motion representations from these Atomic Actions, capturing their spatiotemporal relationships. In the Recompose phase, we compose local motion representations with a learnable Motion Aggregation Token [MAT] via latent dynamics model learning. Additionally, an adapter bridges local motion and downstream action-specific dynamics to accelerate policy learning. Extensive experiments demonstrate that our method effectively transfers to diverse robotic control and manipulation tasks, significantly improving sample efficiency and performance.
arXiv:2607.00862v1 Announce Type: new Abstract: Large Reasoning Models (LRMs) have achieved remarkable success on complex tasks by leveraging long chain-of-thought (CoT) trajectories, yet they frequently exhibit overthinking on simple queries, resulting in significant token overhead and reduced inference efficiency. However, existing compression methods predominantly apply uniform length reduction or rely on coarse-grained difficulty estimation, often leading to performance degradation on difficult problems. To address this limitation, we propose Confidence-Adaptive Thinking (CAT), a framework that incorporates the model's intrinsic self-certainty signals as confidence into the preference optimization process, which autonomously modulates reasoning lengths based on problem difficulty. Experimental results show that CAT consistently outperforms state-of-the-art baselines on reasoning accuracy across multiple benchmarks on different base models. Our work enables LRMs to effectively compress confident responses while deliberating on uncertain ones, offering a potentially robust solution for balancing accuracy and latency in practical industrial scenarios.
arXiv:2607.00442v1 Announce Type: new Abstract: Reinforcement learning (RL) for quadruped locomotion commonly depends on fixed, hand-crafted, and Markovian reward functions that limit both interpretability of learned policies and lack explicit control over gait behaviors. We introduce a framework where distinct gaits are specified using parameterized constraints expressed in Signal Temporal Logic (STL). These include safety bounds, gait synchronization constraints, command tracking, and actuation bounds. From these specifications, we develop a reward shaping mechanism that provides learning agents a dense, continuous reward landscape that encodes desired behavior. We define parametric STL templates for three speed regimes (walking-trot, trot, bound), calibrate their parameters from reference rollouts, and compute rewards from using smooth approximations of STL robustness over the rollouts. The generated rewards can be used to provide shaped gradients compatible with Proximal Policy Optimization (PPO). We instantiate the approach on Google's Barkour quadruped robot in MuJoCo XLA (MJX). We use parallelization within the simulator to improve training speeds and use domain randomization to robustify learned policies. We show that compared to a baseline of hand-crafted rewards, the STL-shaped rewards yield tighter velocity tracking and more stable training. Videos can be found on our project website: https://stl-locomotion.github.io/.
arXiv:2607.00272v1 Announce Type: new Abstract: Traditional robot programming is challenging: it requires orchestrating multimodal perception, managing physical contact dynamics, and handling diverse configurations and execution failures. We introduce ASPIRE (Agentic Skill Programming through Iterative Robot Exploration), a continual learning system that autonomously writes and refines robot control programs in a code-as-policy paradigm while compounding experience into a reusable skill library. ASPIRE discovers skills that persist across tasks, simulation and real-world settings, and embodiments. It operates in an open-ended loop with three components: (1) a closed-loop robot execution engine that exposes fine-grained multimodal traces, enabling autonomous failure diagnosis, repair synthesis, and validation; (2) a continually expanding skill library that distills validated fixes into reusable, transferable knowledge; and (3) evolutionary search that generates diverse task sequences and control programs to explore beyond single-trajectory refinement. ASPIRE surpasses prior methods by up to 77% on LIBERO-Pro manipulation under perturbation, 72% on Robosuite bimanual handover, and 32% on BEHAVIOR-1K long-horizon household tasks. Its accumulated library also enables zero-shot generalization to unseen long-horizon tasks: on LIBERO-Pro Long, ASPIRE achieves 31% success versus 4% for prior methods despite their use of test-time reasoning and retries. Finally, simulation-discovered skills provide initial evidence of sim-to-real transfer, substantially reducing real-robot programming effort across different embodiments and robot APIs.
arXiv:2607.00148v1 Announce Type: new Abstract: Learning predictive models of the world enables robotic control through planning, potentially allowing robots to improvise solutions on new tasks. However, large video-based dynamics models lack explicit 3D spatial structure and suffer from geometrically inconsistent long-term rollouts with compounding errors. Emerging 3D dynamics models based on partial point clouds improve geometric consistency but remain sensitive to occlusions and accumulated prediction drift. To address these challenges, we present 3D Point World Models (3DPWM) - a task-agnostic world model that operates entirely in 3D space by first completing partial point clouds and then learning action-conditioned dynamics in this completed 3D scene. By operating on completed geometry, 3DPWM enables reliable long-horizon rollouts and more accurate cost evaluation for model-based planning while supporting adaptation to new tasks. Experiments across different robotic embodiments and tabletop manipulation benchmarks demonstrate that 3DPWM achieves significantly more reliable long-horizon rollouts (100-300+ steps), supports both open-loop and closed-loop planning, and enables successful sim-to-real transfer.
arXiv:2607.00160v1 Announce Type: new Abstract: Modular reconfigurable robotic systems provide a scalable solution for cooperative surface operations in future lunar missions. However, cooperative cargo transportation remains challenging due to morphology-dependent topology changes, strong payload-induced coupling, long-horizon decision making, and safety constraints. This paper proposes a phase-decomposed reinforcement learning framework for cooperative cargo transport with distributed robotic units. The task is decomposed into lifting, transportation, and placement, each optimized with a dedicated joint-state policy capturing inter-agent coupling. Centralized training promotes stable convergence, while deployment uses onboard proprioception for control and OptiTrack motion capture for ground-truth evaluation and post-processed metrics. A deterministic phase controller expressed in Markov state representation regulates transitions between stages, and a failure-sensitive synchronization mechanism ensures coordinated progression and safety-aware halting during real-world execution. The framework is evaluated in simulation and through controlled field experiments at a JAXA space exploration test facility. Results demonstrate reliable cooperative transport across all stages in both simulation and hardware experiments.
arXiv:2607.00166v1 Announce Type: new Abstract: Multireference behavior in molecules often arises when a small gap between frontier orbitals results in mixing of closed and open-shell configurations. Standard multireference diagnostics of this regime usually rely on correlated wavefunctions, natural-orbital occupations, or reduced density matrices. Here, we examine a complementary, symmetry-based criterion for a model system. For a time-reversal-invariant Hamiltonian, a symmetry-preserving, closed-shell Slater determinant must transform as the trivial irreducible representation of its point group. Therefore, a nontrivial, many-electron irreducible representation excludes such a description. We compare two pathways within the same model to demonstrate this. Along the control pathway, the frontier orbitals remain separated and the ground state retains a trivial irreducible representation over the weak-to-intermediate interaction regime. Along the obstructed pathway, a high-symmetry point produces a frontier-orbital degeneracy, resulting in a singlet ground state with two-configuration character and a nontrivial irreducible representation. Exact diagonalization, a two-state effective model, and the Frobenius norm of the two-particle cumulant provide a consistent picture in this regime, demonstrating that irreducible representations can serve as a low-cost diagnostic of multireference character in diradicaloid models. While symmetry is not a quantitative measure of correlation strength, it does offer a computationally inexpensive screening tool to identify obstructions to a single-reference description.
arXiv:2607.00170v1 Announce Type: new Abstract: Thermodynamic computing devices based on the Ising model show great promise for low-power AI inference and edge computing, but scalable methods for training large models for such hardware remain limited. Prior theory shows that the time-averaged behavior of high-temperature Gibbs-sampled Ising systems can implement feed-forward neural inference. We turn this theoretical correspondence into a scalable and purely backpropagation-based algorithm for training deep convolutional networks for thermodynamic inference on Ising machine hardware. Our image classification models achieve accuracies of 94.9% on CIFAR-10 and 76.0% on CIFAR-100 under binary Gibbs sampling. We then develop and experimentally validate a mathematical theory relating inference cost to accuracy and controlling autocorrelation times. Subsequently, we calculate asymptotic results showing that inference cost is bounded by a well-controlled tradeoff with performance and exhibit algorithms for computing optimal inference schedules. Finally, we discuss implications for hardware development and the future of high-temperature thermodynamic AI models.
arXiv:2606.31983v2 Announce Type: replace Abstract: There has been extensive research on automating and scaling data cleaning, i.e., the detection and correction of erroneous values in tabular data. Yet, existing approaches often perform well only within controlled environments. One of the major bottlenecks in data cleaning research is the lack of real-world datasets. In this paper, we address this gap by providing a large, dirty dataset with postal entries and their corresponding ground truth. We discuss the design decisions and challenges for obtaining the dataset. We demonstrate the limitations of existing cleaning approaches when faced with our proposed datasets and derive guidelines for future research.
arXiv:2607.00252v1 Announce Type: new Abstract: We present an algorithm for the group distributionally robust (GDR) least squares problem. Given $m$ groups, a parameter vector in $\mathbb{R}^d$, and stacked design matrices and responses $\mathbf{A}$ and $\mathbf{b}$, our algorithm obtains a $(1+\varepsilon)$-multiplicative optimal solution using $\widetilde{O}(\min\{\mathsf{rank}(\mathbf{A}),m\}^{1/3}\varepsilon^{-2/3})$ linear-system-solves of matrices of the form $\mathbf{A}^{\top}\mathbf{B}\mathbf{A}$ for block-diagonal $\mathbf{B}$. Our technical methods follow from a recent geometric construction, block Lewis weights, that relates the empirical GDR problem to a carefully chosen least squares problem and an application of accelerated proximal methods. Our algorithm improves over known interior point methods for moderate accuracy regimes and matches the state-of-the-art guarantees for the special case of $\ell_{\infty}$ regression. We also give algorithms that smoothly interpolate between minimizing the average least squares loss and the distributionally robust loss.
arXiv:2607.00254v1 Announce Type: new Abstract: Many modern AI workflows, ranging from LLM post-training pipelines to agentic reasoning tasks, can be expressed as declarative queries whose expensive predicate is evaluated by a large model or reward function. We propose a query-centric formulation of these workflows and show that classical database techniques, namely approximate query processing (AQP) and proxy-model (PM) based filtering, can substantially reduce the number of expensive model invocations without requiring changes to the underlying models or pipelines. Our first strategy treats the workflow as an online aggregation problem: it progressively samples records, maintains a running aggregate estimate with a confidence interval, and terminates early once the interval stabilizes, accepting the estimate when it falls within a user-specified error bound. Our second strategy trains a lightweight, CPU-resident decision tree on a small set of oracle-labeled examples and uses it to pre-filter records whose outcome can be predicted with high confidence, routing only uncertain records to the expensive model. We evaluate both strategies on TPC-DS aggregate queries and on real LLM post-training pipelines including math reasoning, general instruction following, and code generation. On TPC-DS, Strategy AQP keeps aggregate error under 10% while reaching its adaptive stopping point at 10-15% of oracle calls under balanced distributions, an 85-90% reduction, and Strategy PM reduces oracle calls by 60-70%. On LLM pipelines, Strategy AQP reaches its adaptive stopping point at 20-50% of oracle calls with less than 5% accuracy loss on the structured math and code tasks; open-ended instruction following, scored by a reward model, shows a larger but bounded reduction. Strategy PM reduces reward-model scoring time by up to 19x on structured tasks with less than 10% accuracy loss.
arXiv:2607.00018v1 Announce Type: new Abstract: Media coverage of armed conflict is deeply asymmetric: we document a 224$\times$ gap between the most and least covered conflict zones in English-language media across 22 countries (2020--2026). We evaluate zero-shot conflict escalation forecasting across all 22 countries on a 660-case held-out test set, comparing Llama-3.3-70B and GPT-4o against three structured baselines. The central finding is not a performance gradient but a qualitative failure: LLMs do not forecast conflict -- they categorize it. Llama predicts escalation on every under-covered case, matching the trivial Always-YES baseline to three decimals; GPT-4o predicts NO on every over-covered case, missing all five actual escalation events. A logistic regression using only eleven observation-window features with \emph{no country information} achieves F1~=~0.402, outperforming both LLMs in every measurable tier. This failure cannot be resolved at inference time: adding structured ACLED evidence degrades performance on under-covered zones (GPT-4o F1: 0.323~$\to$~0.168) and falls below LR by a factor of 2.4. The bottleneck is not data availability but the LLM's interpretation of temporal signal under a country-categorical prior. Under-covered populations receive not just less accurate AI, but qualitatively different AI that cannot distinguish stable from escalating periods. We call for coverage-stratified benchmarking, conflict NLP datasets for under-covered zones, and training data documentation standards for geographic conflict representation.
arXiv:2607.00039v1 Announce Type: new Abstract: The emergence of Software-Defined Vehicles represents a fundamental shift in automotive design, prioritizing software-centric architectures over traditional hardware-driven models. SDVs require modularity, interoperability, real-time processing, and over-the-air update capabilities throughout the vehicle lifecycle. However, current vehicle systems, characterized by tightly coupled software and hardware, struggle to meet these demands due to their complexity and heterogeneity. A critical first step toward enabling SDVs is the decoupling of software from hardware, which can be facilitated through a robust Hardware Abstraction Layer. While existing HALs offer hardware independence and standardized interfaces, their applicability and effectiveness in SDV contexts remain uncertain. This paper systematically evaluates current automotive HALs and explores HAL mechanisms from non-automotive domains, including smartphones, networking, and industrial automation, to extract cross-domain insights relevant to SDV development. A criteria-driven evaluation framework is developed to assess HALs against SDV-specific needs. Findings reveal that while middleware-based HALs offer portability and modularity, hypervisor-based approaches better support safety, OTA readiness, and hardware efficiency. Limitations in both approaches are identified, prompting recommendations for a hybrid HAL design that integrates hypervisor isolation with middleware standardization. This paper contributes to the ongoing developments in automotive software architecture by offering insights into the applicability and effectiveness of current HAL strategies. It provides actionable guidance for designing flexible, scalable, and future-ready HALs to support SDVs across their lifecycle.
arXiv:2603.20467v2 Announce Type: replace-cross Abstract: Stochastic differential equations (SDEs), which serve as the governing equations for dynamical systems in a broad range of applications, can become cost-prohibitive for numerical simulation at scales necessary for quantifying key properties. Surrogate models of the drift function of an SDE, learned from data of the high-fidelity system, are routinely used to increase the efficiency of simulation and prediction of properties. However, standard choices of loss function for learning the surrogate model fail to provide error guarantees in certain path-dependent observables, such as transition times. This paper introduces an error bound for path-space observables and employs it as a novel variational loss for the goal-oriented learning of the drift function of a SDE. We show the error bound holds for a broad class of observables, including mean first hitting times on unbounded time domains. We derive an analytical gradient of the goal-oriented loss by leveraging the formula for Fr\'echet derivatives of expected path functionals, which remains tractable for implementation in stochastic gradient descent schemes. We demonstrate that surrogate models of overdamped Langevin systems developed via goal-oriented learning achieve improved accuracy in predicting the statistics of a first hitting time observable and robustness to distributional shift in the data.
arXiv:2604.18926v2 Announce Type: replace-cross Abstract: This study presents a mathematical optimization framework and analysis to inform practical long-term investment planning in Puerto Rico's electric power system. We utilize a high-resolution capacity expansion planning model to identify least-cost generation and storage investments that improve reliability. The model co-optimizes new investments with thermal retirements and includes detailed dispatch, unit commitment, fuel selection, storage operation, engineering limits, system constraints, fuel supply limits, and load balance. Key advances over prior studies on Puerto Rico's system include: (i) Nodal transmission representation at 38 kV and above; (ii) hourly chronological simulation for representative days; (iii) explicit unit commitment for existing and new thermal units with realistic ramping, minimum up and down times, and startup costs; (iv) system-wide fuel supply constraints; and (v) operational scenarios reflecting load variability, renewable availability, and high forced outage rates in legacy units. Using data from LUMA, the Puerto Rico Electric Power Authority (PREPA), U.S. Department of Energy, and public sources, the study builds representative Puerto Rico systems for 2024 and 2030, with the latter including planned generation and storage projects. It tests scenarios with different future load levels, fuel supply assumptions, planned additions, and allowed technologies. Under the study assumptions, least-cost portfolios that meet reliability targets require about 1.5 GW or more of new H-class combined cycle capacity, in addition to planned projects. These additions mainly replace unreliable legacy thermal units rather than serve new load. The new CC investments eliminate modeled load shedding in the bulk system and restore a robust reserve margin, even under stressed load and outage conditions.
arXiv:2607.00267v1 Announce Type: new Abstract: A central goal of science is to produce valid explanations of complex systems: high-level causal accounts that faithfully reflect the behavior of lower-level mechanisms. Yet no consensus exists on how to measure whether a proposed high-level explanation is actually valid. We introduce a benchmark of ten complex systems spanning both discrete and continuous state spaces, as well as static and dynamical regimes, each equipped with consensual ground-truth causal explanations and invalid contrastive conditions. Within a unified causal abstraction framework, we systematically evaluate over thirty candidate metrics drawn from observational, functional, information-theoretic, and causal families. Our results show that only the latter reliably discriminates valid from invalid abstractions, and only when incorporating faithfulness testing over unmapped variables. Building on these findings, we introduce the Causal Abstraction Error (CAE), a continuous validity metric with an explicit faithfulness test, which passes all discrimination tests across every system and can converge with as few as 30 sampled interventions. We offer it as a general-purpose metric for the discovery and validation of high-level explanations.
arXiv:2607.00935v1 Announce Type: new Abstract: High adoption of electric vehicles (EVs) can overload distribution transformers when charging requests with heterogeneous departure deadlines compete for limited capacity. Most existing coordination schemes enforce hard deadlines and strict transformer limits, implicitly assuming feasibility and failing under severe congestion. We propose a deadline-aware EV charging framework that explicitly trades off transformer thermal aging and charging service quality under capacity-constrained operation. We model transformer stress using a convex aging proxy and soften charging deadlines via penalty-weighted unmet energy at departure. We further develop a low-complexity online charging policy that prioritizes EVs based on a marginal-cost-aware urgency index. We demonstrate through case studies under increasing EV penetration that the proposed approach reduces transformer aging while preferentially allocating limited capacity to time-critical EVs, closely approximating offline benchmark performance using only real-time information.
arXiv:2607.00302v1 Announce Type: new Abstract: Touch supplies the physical grounding needed to perceive intrinsic material properties, such as friction and compliance, that vision alone often cannot resolve. Recent efforts for equipping multimodal LLMs with this tactile sense, however, expose a zero-sum trade-off: the limited parameter budget of compact models forces a choice between acquiring the new sensory modality and preserving the established vision-language reasoning. We present Splash, a mask-isolated tactile alignment learning framework for MLLMs. Splash quantifies the significance of each pretrained parameter, and partitions the parameter space into a dormant and critical subspace. While the frozen critical subspace acts as a stable anchor to safeguard general visual knowledge, Splash updates the isolated dormant subspace to internalize tactile alignment towards LLMs. This selective, non-destructive expansion effectively prevents catastrophic forgetting and ensures non-destructive modality expansion. Extensive experiments show that Splash effectively achieves tactile reasoning without additional inference overhead in the LLM part, demonstrating state-of-the-art performance on visuo-tactile benchmarks, including SSVTP, TVL, and TacQuad, while preserving its original general-purpose capabilities.
arXiv:2607.00304v1 Announce Type: new Abstract: The bias-reliability tradeoff conjectures that LLM evaluation systems are constrained in (gamma, H, CV) space, where evaluator coupling (gamma), strategy diversity (H), and small-sample measurement reliability (CV(N)) cannot be simultaneously optimized at fixed sample size N. Prior evidence rests on n=5 conditions with complete metrics from a single study. We expand the empirical base to 11 conditions, measuring gamma and H for all 11 (nine with valid weight vectors) and CV(N=5) for seven with sufficient seeds (N >= 5). Five conditions provide the complete (gamma, H, CV) triple. The data confirm the trade-off: conditions with low evaluator coupling (gamma < 0.2) exhibit high measurement noise (CV(N=5) > 1.0), while conditions with strong coupling (gamma > 0.9) achieve low noise (CV(N=5) < 0.16). The correlation r(H, gamma) = -0.989 (n=5, excluding GPT-4o conditions) confirms that evaluator coupling suppresses strategy diversity. Four GPT-4o conditions show gamma=0.000 and H=1.000 across all seeds -- a pattern we attribute to version drift in the June 2026 GPT-4o API. No condition occupies the region {gamma < 0.2, CV(N=5) < 0.3}. We release all per-condition metrics as a standardized benchmark dataset for evaluator comparison.
arXiv:2607.00218v1 Announce Type: new Abstract: Vision-language models (VLMs) are now proposed as runtime safety guards for embodied agents in homes and factories. A deployable guard must catch genuinely unsafe situations while avoiding unnecessary intervention on routine but superficially alarming activity, a distinction that binary safety benchmarks obscure. We introduce EgoSafetyBench, an egocentric video benchmark of 1,200 robot-view scenarios annotated at half-second granularity, to evaluate VLMs as streaming guards across two tracks. The situational track (800 scenarios) spans four families, from routine and safe-but-suspicious scenes to obvious and contextual hazards. The visual-channel track (400 scenarios) targets in-scene text-a sign, sticker, or label visible in the scene-that can misrepresent the physical situation, pairing each misleading sign with a truthful version to test both whether a guard flags the text as misleading and whether the text corrupts its physical-safety judgment. Both tracks use contrastive ladders: near-identical scenarios differing only in a single visible deciding cue, so a correct call must hinge on that cue rather than the overall scene type. We evaluate ten open- and closed-source VLMs. We find that while guards reliably recognize videos containing hazards, they often miss specific hazardous moments, particularly contextual hazards. Furthermore, misleading in-scene signs degrade all tested guards: vulnerable models miss up to a third of hazards, while robust models over-intervene on safe content. Matched controls reveal that apparent safety robustness often reflects indiscriminate alarming rather than true physical reasoning.
arXiv:2607.00313v1 Announce Type: new Abstract: Let $\mathsf{E}=\mathbb F_q[x]/(\Gamma)$ be an algebraic extension of degree $n$ over the finite field $\mathbb F_q$, given by a $\Gamma\in\mathbb F_q[x]$ monic and irreducible. It is classical that any such $\mathsf{E}$ contains an element $\beta\in\mathsf{E}$ that is normal over $\mathbb F_q$, i.e., the conjugates $\beta,\beta^q,\ldots,\beta^{q^{n-1}}$ form an $\mathbb F_q$-basis of $\mathsf{E}$. In this paper we give a deterministic algorithm which finds such a normal element using $O_\epsilon((n^2\log q)^{1+\epsilon})+O\,\tilde{}\,(n\log^2 q)$ bit operations, for any $\epsilon>0$. The algorithm works by showing that, for a parameter $t\in\mathbb F_q$, the element $\beta_t=(\theta-t)^{-1}$ is normal except for at most $n(n-1)$ values of $t$. This is established by constructing a "cleared Moore" circulant matrix over $\mathbb F_{q^n}[\mathcal T]$, whose determinant degree at most $n(n-1)$, such that $\beta_t$ is normal if and only the determinant is non-zero at $t\in\mathbb F_q$. For faster computation over the base field, we replace this by an equivalent trace Gram circulant matrix over $\mathbb F_q[\mathcal T]$. A main algorithmic contribution is a fast determinant algorithm for circulant matrices of polynomials, which uses triangular set projection and modular composition techniques to achieve a near-linear cost. Given an $n\times n$ circulant matrix over $\mathbb F_q[t]$ whose entries have degree at most $m>0$, we show how to compute its determinant deterministically with $O_\epsilon((nm\log q)^{1+\epsilon})$ bit operations. We complete the solution by showing how to extend this to finite fields of size less than $n(n-1)$, through an embedding in a low-degree extension field, at poly-logarithmic additional cost.
arXiv:2607.00315v1 Announce Type: new Abstract: While the Church-Turing thesis asserts that effective calculability explicates to sets decidable by a Turing machine, the Cobham-Edmonds thesis asserts that feasible computation explicates to the complexity class $\mathsf{P}$, those decidable by a polynomial-time bounded Turing machine. The Church-Turing thesis has been placed under rigorous scrutiny and has several convincing arguments in its favor, but the Cobham-Edmonds thesis has not undergone a similar examination. Many of the arguments in its favor simply suggest that $\mathsf{P}$ is a useful assumption, rather than a necessary target. This paper presents analogous arguments in favor of the Cobham-Edmonds thesis.
arXiv:2607.00857v1 Announce Type: new Abstract: One of the three key hypotheses of fine-grained complexity asserts that computing All-Pairs Shortest Paths (APSP) requires cubic time, up to subpolynomial factors, in the worst case. We initiate the study of APSP in the paradigm of algorithms with predictions, also known as learning-augmented algorithms. We propose an APSP algorithm that takes as additional input a \emph{prediction} (e.g., given by a model learned from similar instances seen in the past) consisting of sets of vertices causing the shortest \emph{detour} for each pair of vertices. The algorithm runs in time $\mathcal{O}(n^{2.83} + \eta n)$, where $\eta$ denotes the \emph{prediction error} defined as the number of pairs of vertices for which, informally speaking, the prediction was not sufficient to compute and certify optimality of the shortest path length. This is already subcubic when the prediction error is (polynomially) smaller than its maximum possible values $n^2$, i.e., whenever the prediction is at least slightly better than terrible. We build on the co-nondeterministic algorithm for the Exact Triangle problem by Chan, Vassilevska Williams, and Xu (STOC 2023), essentially enabling this algorithm to detect mistakes in the nondeterministic certificate and recover from them. Our result constitutes the first necessary step towards designing learning-augmented algorithms for problems with known fine-grained lower bounds conditioned on the APSP Hypothesis.