Papers

World Action Models (WAMs) couple predictive world modeling with action generation, allowing anticipated future states to guide agent behavior. Although WAMs are rapidly advancing embodied AI, general-purpose counterparts remain largely unexplored in games. Existing game-oriented approaches often combine action-conditioned world models with external policies and reward functions to realize WAM-like decision-making, yet they operate mainly in 2D visual observation space and do not instantiate persistent 3D geometry. Extending this paradigm to 3D games introduces a distinct challenge. In autonomous driving and robotics, the physical environment exists independently of the model, providing a persistent 3D world in which selected actions can be executed. Games have no such external substrate; the virtual world itself must be instantiated. Most playable games require a persistent and navigable space, while 3D games additionally require explicit geometry that supports movement and interaction. Action-conditioned video rollouts provide visual observations but not this spatial representation. We present Valerant, a training-free framework that transforms a pretrained action-conditioned world model into a WAM for exploring and constructing 3D game maps. By coupling predictive visual rollouts with SLAM-based spatial reconstruction and exploration-driven action selection, Valerant progressively transforms a single image into a persistent 3D game map. This framework extends WAM-based interaction beyond 2D visual simulation and offers a new approach to reducing manual effort in 3D game-map creation.

Vision-language models (VLMs) show promise for agricultural diagnosis but fail unpredictably in zero-shot settings. We test whether this stems from poor visual features or an inability to verbalize what the model already perceives, using a benchmark spanning classification and detection drawn from 118 AgML classification and 54 detection datasets. Linear probing shows agricultural features are linearly separable in frozen VLM vision towers, with Gemma-4-31B's encoder reaching F1-macro of 0.60–0.78 across tasks and approaching the vision-only DINOv3 backbone, indicating the bottleneck lies downstream of perception. We then test whether test-time scaling, an iterative in-context skill-memory loop, can close this gap without updating weights. Gains are modest and inconsistent. Skill-memory improves classification accuracy by up to 3.1 points on three of four tasks while regressing on species, with gains concentrated on classes seen during optimization and unseen classes regressing. Detection F1 is effectively unchanged (0.412 to 0.421), though an oracle filtering condition reaches 0.432 F1, revealing real headroom unconstrained prompting fails to surface. These results indicate VLMs see more agricultural features than they report, but test-time scaling alone fails to surface that hidden parametric knowledge, motivating agentic approaches grounded in external, retrievable knowledge.

Choosing a recovery process for scale-up requires connecting laboratory results with product requirements, process costs, and scale effects. We analyze records from Pacific Northwest National Laboratory's Computer Intelligence for Critical Element Recovery and Optimization (CICERO) workflow for autonomous selective precipitation. Active learning uses prior results to choose experiments. In a conditional retrospective benchmark with fitted models and recycled neodymium-iron-boron (NdFeB) magnet records, active learning finds the best recorded result with fewer experiments than nonadaptive space filling. Enrichment is the selected rare-earth-to-iron ratio relative to that in the feed. Adaptive policies reach the recorded enrichment maximum by 16 to 24 wells (individual experiments), versus 48. Our two-stage reconstruction ties two adaptive alternatives at 16 wells. Conditional analyses of recycled samarium-cobalt (SmCo) magnets show a Round 2 tradeoff between purity and nominal yield, the recovery fraction calculated from an assumed starting amount - NdFeB Round 1 routes differ in enrichment. Rankings for produced water from oil and gas extraction depend on phase and dilution assumptions requiring confirmation. We propose choosing batches by their expected reduction in downstream Bayes risk: the minimum expected loss among available process decisions under current beliefs. In exploratory simulations, a hybrid that filters candidates has lower estimated loss than the implemented joint search across routes and conditions. Differences involving the synthetic two-stage policy are small relative to estimation uncertainty. We outline a pre-registered prospective test under a shared loss and logging standard, requiring clarified measurements and records, a defined process decision and relevant outputs, credible economic inputs, and validation at the intended scale.

While autonomous agents have made significant strides in "deep research" by iteratively navigating the open web to synthesize information, real-world problem-solving is rarely confined to a single environment. Complex analytical tasks inherently require agents to weave together evidence from both ambiguous unstructured text (e.g., the open web) and highly precise structured data (e.g., relational databases). However, existing benchmarks evaluate these modalities in isolation, failing to capture the critical "handoff" - the ability to preserve constraints when moving evidence between systems. We introduce HybridDeepResearch, to our knowledge the first deep-research benchmark that requires both web search and SQL to form a complete, verifiable answer. The benchmark contains 380 tool-dependent tasks grounded in LiveSQLBench-Base-Lite databases and public web corpora, validated through automated checks and human review, and covering three reasoning patterns: SQL2S, S2SQL, and Parallel. Evaluations across proprietary and open-weight models under various agentic scaffolds reveal that even state-of-the-art models like GLM-5.2, Claude-Sonnet-4.6 and GPT-5 achieve only about 50-54% Pass@8 on the hard subset. Notably, results show that directional reasoning is substantially more difficult than parallel intersection, highlighting that bridging structured and unstructured information spaces without losing constraints remains a major open challenge for agentic systems. Code and datasets are publicly available at GitHub (https://github.com/Snowflake-AI-Research/HybridDeepResearch) and Hugging Face (https://huggingface.co/datasets/Snowflake/HybridDeepResearch).

Near-field localization is a promising technique for high-resolution multi-user positioning in future wireless systems, but its performance is often degraded by scattering-induced coherent propagation. Existing near-field localization methods, which require separate parameter estimation and path/source association, suffer from high computation overhead and accumulated errors, and usually do not provide any guarantee on reliability. In this paper, we propose MUSIC-Net, an end-to-end near-field positioning deep learning (DL) framework informed by two-stage MUltiple SIgnal Classification (MUSIC) in mixed line-of-sight (LoS) and non-LoS (NLoS) multi-path scenarios, which embeds the two-stage MUSIC objects into training to isolate the LoS-related signal subspace and to identify a surrogate distance. The proposed framework directly recovers multi-user positions without the need for involved NLoS parameter estimation or path/source association. Furthermore, we introduce split conformal prediction (SCP) to move beyond point-estimation-based positioning towards statistically guaranteed (confidence) set estimation for all users. Numerical results show that the proposed MUSIC-Net achieves lower mean positioning error (MPER) than existing benchmarks and yields tighter SCP-calibrated prediction regions, demonstrating both accurate LoS localization and efficient uncertainty quantification (UQ) in coherent multi-path environments.

Agentic AI frameworks let a language model plan, keep memory, and call tools that reach real files, mail, and services. Most of these agents also read images, which gives an attacker a way to put text into the agent's context without going through the user. We present MMPIBench, a reproducible benchmark that measures what happens next. It delivers a fixed set of attacks through six visual carriers (OCR text, overlays, EXIF metadata, QR codes, fake interfaces, and hybrids) and records how far each injected instruction travels through the agent, from perception through planning to the tool call. Across 720 runs covering six frameworks, five foundation models, six carriers, and four attacker objectives, attacks complete in approximately 1% of runs but are attempted in 12.8%, and the gap is closed almost entirely at the planning step, where the model reads the injected instruction and declines to act on it. The model matters far more than the framework for whether an instruction is acted on. One model never attempts an attack and recognizes the injection in 59.7% of runs, while two others attempt in 23.6%. We then extend the benchmark to audio, the only other raw perceptual channel current frontier models accept. Only two of the five models ingest audio and only three of the six frameworks deliver it, but where the signal arrives the attack completes in 49% of cells, and in 75% for one model. Reporting completion alone therefore understates exposure, and perceptual channels beyond vision are narrower but much less defended.

In the past decade, Bayesian optimization (BO) has emerged as a powerful and adaptable framework for automatic controller tuning and robot learning. This article offers a comprehensive overview of the state-of-the-art in BO, designed to support both researchers and practitioners in understanding recent advancements, practical applications, and future research directions. We begin by adopting a practitioner's perspective, illustrating how to effectively set up BO through a representative controller tuning example. We position BO within the broader context of learning paradigms, ranging from deep reinforcement learning to data-driven control, and highlight scenarios where BO is most advantageous. Next, we discuss the diverse range of BO methods that have been developed to tackle complex problems and specific applications. This article provides a unified perspective on the current landscape of BO, emphasizing its relevance to control systems and robotics, and it highlights future prospects by identifying key research challenges and promising avenues for advancing BO in the field. This includes addressing a significant gap in the BO landscape: the lack of standardized benchmark problems specifically for control-related applications. To foster future research and ensure rigorous evaluation, we start an effort towards a lightweight benchmark suite for control engineering and robotics. We also present metrics and best practices to facilitate direct comparisons between new BO algorithms and established state-of-the-art methods.

Enterprises managing large X.509 certificate inventories face a prioritization problem: deterministic analysis tools that precisely identify standards violations are indispensable for remediation, but applying them exhaustively across millions of certificates is operationally impractical. We present X-amine509, a two-stage triage system that uses machine learning to rapidly rank certificates by predicted risk and route only the highest-risk items to full deterministic analysis. Certificate risk is quantified as a composite score derived from 177 defect checks grounded in CA/Browser Forum Baseline Requirements, NIST IR 8547/SP 800-57, and cryptographic strength criteria, weighted by security severity across four tiers ranging from cryptographic breaks to minor compliance deviations. We collected 1,027,714 X.509 certificates from Fortune 500, .gov, and .edu domains and scored each using this rubric. On a held-out test set of 201,976 certificates, our best model (Extra Trees) achieves $R^2$ of 0.993 with MAE of 2.26, while Decision Tree scores $R^2$ of 0.986 at 3.7 million certificates per second on a single machine. Ranking quality confirms the triage value: aggregate NDCG exceeds 0.997, and severity-tier classification reports 99.76% accuracy with 98.90% recall on critical-tier defects. Thirteen months later, we retrieved another 571,374 certificates to test our models' durability over time, and the Extra Trees and Decision Tree models maintain MAE below 6.8, $R^2$ of at least 0.915, aggregate NDCG above 0.988, severity-tier accuracy of at least 99.52%, and critical-tier recall of at least 97.03%. Feature importance analysis identifies validity period, Extended Key Usage configuration, negative serial number encoding, and self-signed status as the strongest risk predictors, providing coarse interpretability at the triage stage.

Reconfigurable optical data center networks (RODCNs) have emerged as a promising solution for scaling DCN capacity, yet their scheduling mechanisms remain a performance bottleneck: traffic-oblivious schemes inherently limit throughput, while the state-of-the-art traffic-aware scheme, NegotiaToR, uses single-iteration iSLIP as its scheduling engine, which limits throughput to around 60% and treats all source-destination pairs with equal priority regardless of queue length. We propose QPS-ToR, which replaces NegotiaToR's scheduling logic with SW-QPS, a sliding-window algorithm originally proposed for crossbar scheduling that achieves around 90% throughput with a single low-complexity iteration. QPS-ToR operates within NegotiaToR's existing workflow, requiring only a revision of the scheduling cycle from three-step Request-Grant-Accept (RGA) to two-step Request-Grant (RG) with a sliding window mechanism. In flow-level simulations with 128 ToRs under realistic datacenter workloads, QPS-ToR achieves up to 36% higher throughput and 82% lower flow completion time (FCT) compared to NegotiaToR, and consistently outperforms RotorNet, a representative traffic-oblivious scheme, on the parallel network topology.

This note introduces a general data representation of dynamic systems.

The goal of degree sequence reconstruction is to recover the ordered vector of degrees of an unknown graph from vertex deleted traces, where each vertex is deleted independently with probability $p$. We provide two algorithms for reconstruction; the first uses rejection sampling to reduce the problem to an estimation problem for a mixture distribution. Combined with prior trace reconstruction results, this gives a reconstruction algorithm using $\Exp{\tilde O (n^{1/3})}$ traces, although no sub-exponential time decoder is known. Our other approach involves recovering certain graph invariants, degree moments, that can identify a graphs degree sequence. Extremal polynomial bounds show that $\tilde \Theta(n^{1/2})$ degree moments are necessary and sufficient to reconstruct the degree sequence, which leads to an algorithm with $\Exp{\tilde O(n^{1/2})}$ trace complexity. The same polynomial machinery yields a sub-exponential time decoder for the degree sequence from the moments. Additionally, we give an $O(n^{3})$ upper bound and a $\Omega(n^2)$ lower bound for the trace complexity of recovering the number of edges.

As Vision-Language Models (VLMs) advance toward physical deployment, the focus has remained on action-oriented Embodied AI evaluated on subject-centric consumer video. This overlooks a pervasive class of Physical AI: Infrastructure AI, which relies on fixed cameras for open-loop insights like safety monitoring and operational logging. We introduce VANTAGE-Bench, a benchmark measuring this "Infrastructure AI Gap." It spans three operational domains (Logistics, Transportation, and Smart Spaces), unifies image and video evaluation across semantic, spatial, temporal, and spatio-temporal capabilities, and moves beyond multiple-choice to eight task formulations including dense captioning and spatio-temporal grounding. It adds a single-pass trajectory protocol for Single Object Tracking and, to our knowledge, the first such evaluation on fixed-camera infrastructure video, scored against specialist trackers. Annotation spans three regimes over 3,346 media assets: 3,342 video-task annotations, 4,281 image-grounding annotations, and 27,404 detection boxes. Evaluating 17 models zero-shot, we find the shortfall relative to consumer-centric benchmarks is concentrated, not general. Event verification, referring expressions, and temporal localization fall roughly 9 to 24 points at every model scale, while video question answering stays within 5.3 points of VideoMME and 2D spatial pointing shows no shortfall against BLINK. The temporal pillar is weakest in absolute terms: no system exceeds 55.7 mIoU on temporal localization or 37.3 SODA_c on dense video captioning. On tracking, frontier models come within roughly 5 points of specialist trackers over short horizons but separate as the horizon extends. Open-weight models lead 2D object localization outright, so neither scale nor proprietary access explains the pattern. Data, evaluation harness, and leaderboard: https://vantage-bench.org/

Language models act through tools, yet practical agents face libraries containing thousands of interfaces. We introduce the tool menu as the short, ordered subset of available tools shown to an agent before execution. The agent can call only tools in this menu. Multi-step tasks require the final action and the prerequisite tools that create its inputs in a usable order. Current constructors rank tools by request relevance, which can surface the final action while omitting or delaying less obvious producers. We introduce the state path, a pre-execution route from the observable request state to the desired outcome, and propose State-Path Tool Menu to learn it. Our framework treats the menu as an execution prior over these routes. Its encoder represents which tools can run from the current state, how their outputs satisfy later inputs, and which orders recur in training paths. A retriever covers an executable entry, the missing-input producers, and the final action. A reranker then places producers before consumers. On ToolBench, our menu raises online success from 0.737 to 0.898 and outperforms retrieval, reranking, generation, and routing baselines without changing the agent. The State-Path menu also covers more complete chains with 32 tools than the official list covers with 128, and its success gain persists across executor families with different model capacities. Our code is at https://github.com/Met2348/State-Path.

Recovering metric 3D geometry from monocular images is a fundamental computer vision task, yet current methods remain heavily fragmented by fixed camera model assumptions and inflexible input schemes. We present OmniPoint, a unified framework designed to generalize metric reconstruction across diverse imaging sensors, including pinhole, fisheye, and equirectangular projections, while accommodating varying geometric priors. To overcome projection rigidity, OmniPoint abandons conventional planar depth regression. It instead adopts a decoupled ray and distance representation alongside a decoupled training objective, explicitly separating the camera projection model from the scene structure. To address the severe scarcity of training data for alternative cameras, we introduce a bidirectional augmentation strategy that explicitly bridges labeled perspective data and unlabeled omnidirectional domains in 3D space. Furthermore, to seamlessly integrate optional inputs like camera intrinsics or sparse depth without destabilizing the network through feature distribution shifts, we propose a robust information injection mechanism. This mechanism utilizes learnable input state embeddings to resolve architectural ambiguity and applies vectorized Gaussian smoothing to densify irregular measurements. Extensive experiments demonstrate that OmniPoint achieves state-of-the-art zero-shot performance across multiple benchmarks, establishing a robust new standard for unified monocular 3D reconstruction.

Brain-age prediction models are commonly evaluated by predictive accuracy, yet accurate predictions alone do not establish that a model relies on reproducible or neurobiologically supported mechanisms. Post-hoc explanation methods can expose these mechanisms, but existing workflows typically stop at diagnosis or require correction targets to be specified before model analysis. We propose XAI-Refine, an automated explanation-knowledge loop for brain-age prediction from resting-state functional connectivity. At each iteration, XAI-Refine consolidates complementary post-hoc analyses across repeated training runs into reliable, structured model explanations. It converts each reliable explanation into a neutral neurobiological question, retrieves and verifies relevant literature, and compiles the verified evidence into an admissible set in the same typed explanation space. The target for refinement is defined as the minimal projection of the current model explanation onto the admissible set induced by applicable verified knowledge. This revised explanation is then translated into a differentiable constraint while preserving the originating model variable, measurement operator, and applicable scope. Candidate updates are promoted only when multi-seed validation confirms target-directed explanatory movement, predictive performance remains within a prespecified guardrail, and non-target explanatory drift remains bounded. Experiments on functional-connectivity-based brain-age prediction evaluate predictive performance, explanation reliability, literature alignment, and target-specific model revision, illustrating a structured route from post-hoc analysis to evidence-guided model refinement.

Rate adaptation is a crucial mechanism in IEEE 802.11 networks and next-generation cellular systems. Since the time scale for rate adaptation is typically much shorter than that for user association and scheduling, we investigate a joint design of wireless user association and scheduling and rate adaptation across different time scales to maximize cumulative network throughput while ensuring desired fairness among users. We develop a MaxWeight-type user association and scheduling algorithm that integrates virtual queues – tracking each user's scheduling debt to maintain fairness – and Upper Confidence Bound (UCB) estimates in its weight measure. Each selected user then employs the UCB algorithm for rate adaptation on a short time scale. Our theoretical findings reveal that the proposed algorithm achieves cumulative regret that grows with the square root of the time horizon up to a logarithmic factor and results in zero cumulative fairness violation after a certain number of time frames. Furthermore, since the MaxWeight-type algorithm involves evaluating all the feasible schedules that can be exponential to the number of users due to the interference constraints, leading to high computational complexity, we introduce a low-complexity alternative utilizing the so-called pick-and-compare (PC) approach. We demonstrate the effectiveness of both algorithms through simulations based on real-world data traces.

Automated vehicles rely on millions of lines of safety-critical software, yet general-purpose analyzers do not understand which code can affect vehicle motion. This study asks whether large language models (LLMs) with explicit automated-vehicle (AV) security knowledge improve weakness detection beyond rule-based tools. We developed an AV vulnerability taxonomy with 18 weakness classes from vulnerability records, security advisories, and AV-security literature, and integrated it into LLM-based Security Analysis for Automated Vehicles (LLMSec-AV). Evaluated on Autoware, the framework decomposed 770 translation units into 4,673 functions and analyzed 161 functions under four prompting conditions involving taxonomy context, retrieval from 374 prior disclosures, and multi-step analysis. Findings were compared with 46 weakness locations mined from upstream fixes and a flag-volume-matched permutation baseline. CodeQL, Semgrep, cppcheck, and the Clang Static Analyzer evaluated the same code, with AV-specific rules added to CodeQL and Semgrep. Generated fuzzing harnesses were tested using AFL++ and sanitizers. LLM conditions recovered up to 76% of the 46 known weakness locations, outperforming conventional analyzers. CodeQL, Semgrep, and the Clang Static Analyzer matched none, while cppcheck matched one despite 1,301 alerts. Unaided prompting achieved similar detection performance, showing that the taxonomy did not drive recall. However, taxonomy context increased the share of findings assigned to a weakness class from near zero to over 80%, improving interpretability and triage. Six of the 18 classes could not be directly represented as static-analysis rules. LLMSec-AV introduces an AV-specific, machine-readable vulnerability taxonomy for weakness discovery and shows that LLMs can complement conventional analyzers by identifying and organizing safety-relevant findings in real AV software.

This paper addresses safe coordination of scalar agents with constrained input dynamics over fixed rooted digraphs. We propose a networked Admissibility-Preserving Control (APC) architecture for a common moving output corridor under heterogeneous asymmetric input bounds, wherein the Admissibility-Preserving Input Realization (APIR) generates the physical inputs. In logarithmic barrier coordinates, the closed-loop dynamics form an exact cascade of exponentially decaying realization errors and directed consensus dynamics. We derive direction-specific compatibility conditions that guarantee uniform corridor and input margins, bounded commands, unique complete solutions, and exponential consensus for compact initial sets. The closed-form expression for the agreement value quantifies the effect of the initial APIR mismatch. Under the corresponding compatibility conditions, partial pinning guarantees safe reference tracking whenever the root component contains at least one informed agent. Simulations illustrate the coordination guarantees.

Livestock population data disaggregated by age, sex, and production are important inputs to calculations and models that inform our understanding of global health, yet these data are fragmented across disparate sources. Bridging data siloes to improve the findability of data requires interoperability. Conventional approaches to improving the findability and interoperability of data include indexing standardized metadata. However, creating metadata is time and resource-intensive and is often difficult in domains such as livestock, which lack standards that address the needs of broad user groups. When metadata exist, they typically need to be standardized against a pre-existing vocabulary, ontology, or thesaurus, requiring a technique known as `crosswalking'. To overcome issues in the absence of metadata, the lack of standards, and the resource-intensive solutions that currently exist, this study uses a bottom-up approach. By leveraging real-world reporting categories in datasets, the composition and semantics of terms already present in the data were extracted and analyzed. Using cattle data as a pilot, we find the age, sex, and production modifiers present across cattle terms from five datasets from four data sources capture granularity not present in AGROVOC, the largest agricultural vocabulary in the world. We discuss how the composition and semantics of these terms can be used to improve the interoperability and findability of data without first requiring metadata to be generated or standards to be created. Rather than forcing datasets to conform to an existing vocabulary, this approach uses the semantics embedded in terms already present in datasets, allowing systems to make data more discoverable and interoperable while maintaining culturally and dataset-specific terminology.

Unlocking the potential of tiny aerial robots requires order of magnitude improvements in the performance of embedded edge control. In particular, although recent cached model predictive control (MPC) solvers can handle the fast system dynamics and complex constraints required for agile drone flight, their computational demands remain prohibitive for resource-constrained robots, forcing prior implementations to operate at reduced control rates. AccelMPC overcomes this challenge through an end-to-end co-design approach that jointly optimizes the solver algorithm, numerical representation, hardware mapping, and physical integration. AccelMPC pairs a co-designed FPGA-accelerated alternating direction method of multipliers (ADMM)-based MPC solver with a custom 6g PCB, providing high-bandwidth communication for deployment on a 35g Crazyflie. Hardware experiments demonstrate 1 kHz onboard constrained MPC with dynamic obstacles, up to 15.6x faster solve times and 195.4x improvement in energy-delay product over state-of-the-art embedded microcontroller-based solvers, all while scaling to optimization problems with over 20,000 optimization variables and a comparable number of constraints. We release our PCB design files, firmware, and FPGA solver code open source.

Web accessibility auditing requires interpreting diverse requirements and examining interface behavior. Rule-based checks and noninteractive model assessments can miss barriers requiring contextual or interactive evidence. We present an agentic framework that assigns a vision-language agent to each accessibility requirement. Guided by tailored instructions, agents inspect webpages, operate controls, and record evidence supporting their findings. We implement the framework for 40 requirements from the Web Content Accessibility Guidelines (WCAG). To compare detection and cost, we construct a dataset of 250 page-criterion records derived from expert audits across 11 scholarly platforms. Agents recover 67 of 78 reported positive cases (86% recall), compared with 36% for axe-core, a rule-based checker, and 67% for an uncued, noninteractive vision-language model, at lower precision (56%). They recover nine of ten Keyboard and No Keyboard Trap cases missed by both baselines. Together, the framework, implementations, and dataset support automated accessibility auditing grounded in inspectable evidence.

The demand for continuous, high-resolution aerial monitoring is driving interest in High- Altitude Pseudo-Satellites (HAPS), which offer a cost-effective alternative to satellites with greater flexibility. While HAPS platforms provide valuable Earth observation potential, se- lecting an optical sensor requires balancing resolution and payload constraints, given their lightweight structures. We propose a multidisciplinary concept to enhance the Earth obser- vation capabilities of HAPS using a commercially available, weight-efficient distributed camera network, combined with advanced image processing and active vibration control. A super- resolution pipeline is introduced, where images from individual cameras are preprocessed and fed into a super-resolution algorithm. The low structural rigidity of HAPS platforms increases their sensitivity to low-frequency vibrations, which must be mitigated through active control to preserve imaging performance. A nominal H-infinity mixed-sensitivity controller is designed to suppress resonance peaks at camera locations. The methodology is experimentally validated on a simplified wing platform designed for real-time testing. Quantitative evaluation based on nine metrics demonstrates significant image quality improvements, highlighting the effectiveness of H-infinity control in mitigating low-frequency elastic modes and stabilizing the platform for super- resolution. These results show that lightweight commercial sensors, combined with advanced image processing and control, can deliver high-quality imaging for future HAPS remote sensing missions.

Arctic maritime navigation is becoming increasingly important as changing sea-ice conditions expand seasonal accessibility while simultaneously introducing substantial operational, environmental, and community risks. Arctic route planning is inherently a multi-criteria problem: routes that improve vessel safety or efficiency may increase exposure to sea ice, sensitive ecosystems, or nearby communities. Existing routing methods prioritize travel time, fuel use, and navigational risk, often overlooking ecological and community impacts. We introduce a human-in-the-loop, multi-agent GeoAI system for Arctic eco-navigation that integrates operational, physical, ecological, and community-related criteria within a unified routing framework. Multiple specialized agents coordinate geospatial data acquisition and preparation, multi-objective route generation, and skyline-based decision support. The ecological criteria explicitly account for exposure to sensitive areas, including Essential Fish Habitat and seal critical habitat. By considering these ecosystem impacts and potential community burdens while keeping consequential value judgments under human control, the framework supports safer, more transparent, and socially responsible Arctic navigation. Project page and code are publicly available. https://samiraat.github.io/Arctic-Eco-Navigation-Agent/, https://github.com/samiraat/Arctic-Eco-Navigation-Agent

The Holevo barycenter of a quantum channel is the unique output state obtained as the average output of any ensemble achieving the Holevo capacity. Given two quantum channels, the multiplicativity problem asks whether this barycenter tensorizes under parallel composition, namely whether the barycenter of the product channel coincides with the tensor product of the individual barycenters. This question is closely related to the additivity problem for the Holevo capacity: additivity implies tensorization of the Holevo barycenter, while tensorization alone is not sufficient for additivity. Although a construction is known that demonstrates the existence of channels with non-additive Holevo capacity, their corresponding Holevo barycenters still tensorize, leaving open whether multiplicativity might ultimately hold universally. Here, we answer this question in the negative by exhibiting channels for which the Holevo barycenter is not multiplicative under the tensor product of the channel with itself. Moreover, we show that the entropy of the Holevo barycenter is neither universally subadditive nor universally superadditive under tensor product.

Although MMLU is widely adopted as a benchmark for calibrating general AI capabilities, we psychometrically demonstrate that its aggregate score primarily evaluates a model's factual retrieval capacity rather than its reasoning ability. By calibrating item difficulty for 1,000 open-weights language models over 14,042 MMLU test items using Item Response Theory, we show that evaluating both abilities via a single test is inherently flawed. Difficulty is then regressed on a deterministic, text-extractable framework of structural complexity. Applying a joint Wald test with subject-clustered covariances demonstrates that the MMLU conflates fundamentally separable constructs. The mapping from structural complexity to difficulty is not invariant across the benchmark's STEM and non-STEM partitions. This finding has practical consequences. Aggregate leaderboard ranks track non-STEM accuracy more closely than STEM accuracy, so selecting a Top-50 model on the aggregate for a reasoning-intensive deployment displaces roughly 22% of the STEM-appropriate choices. Furthermore, when controlling for the multiple-choice guessing floor natively inside the response model, we find that higher-ability models continue to degrade more steeply under increased reasoning depth. The MMLU aggregate therefore weights retrieval capacity and reasoning stability unequally, inadvertently favoring models optimized for retrieval. We release our deterministic framework as a reproducible auditing instrument and recommend disaggregated reporting.

Discrete black-box optimization is often addressed using approaches such as Sequential Model-Based Optimization (SMBO), which aims to improve sample efficiency by fitting surrogate models that approximate a costly objective function over a discrete search space. In many real-world problems, the set of feasible inputs is often given by logical constraints known in advance. However, existing surrogate modeling techniques generally fail to capture the symbolic rules governing feasibility in discrete input spaces. In this paper, we propose a surrogate modeling approach based on tensor decomposition that captures the structure of discrete search spaces while directly integrating feasibility information. To implement this approach, we formulate surrogate model training as a constrained polynomial optimization problem and solve a relaxed formulation using a differentiable penalty term derived from T-norms. Our experiments on both synthetic and real-world benchmarks, including a pressure vessel design task, demonstrate that the proposed method improves sample efficiency by effectively guiding the search away from infeasible regions.

We present the Living Library, an end-to-end framework for transforming fragmented digital archives into governed, conversational, in-person exhibit experiences. Developed and deployed at the Theodore Roosevelt Presidential Library, the framework comprises four layers: digitization and corpus creation, AI-powered processing, retrieval and reasoning, and an optional embodied conversational interface. The first three layers aggregate a 300,000-record collection, apply OCR and structured metadata enrichment for expert curatorial review, and publish records to a hybrid dense/semantic index. Expert review is conducted through the Archivist App, a curator-facing interface that supports correction of AI-generated transcriptions and metadata. The governed corpus powers both a researcher-facing interface and Talk to TR, a continuously operating exhibit that embodies Theodore Roosevelt as a full-scale digital human within a museum environment. To support live, face-to-face interactions, Cross-Era Analogical Grounding reframes contemporary questions through documented historical parallels, allowing Roosevelt to address present-day topics without inventing facts. Dual-path retrieval and end-to-end streaming keep responses grounded and responsive. Layered watchdogs, visitor-session isolation, automated conversation management, and independently restartable services enable reliable unattended operation for hundreds of visitors. Avatar realism, spatial audio, lighting, staging, and conversational design are developed and evaluated as an integrated experience. Rather than report a controlled benchmark, we describe lessons from operating Talk to TR as a public exhibit and offer a transferable model for transforming archival collections into believable, in-person conversational experiences.

Divergence-based regularization and Sharpness-Aware Minimization (SAM) are two prominent approaches for improving generalization in deep learning, both motivated by robustness to perturbations. However, their relationship has remained largely unexplored. Building on classical second-order expansions of $f$-divergences, we show that the two methods are locally consistent under parameter-space perturbations: both induce curvature-sensitive penalties, with divergence regularization yielding a Fisher-weighted quadratic form and SAM penalizing sharpness through the dominant Hessian eigenvalue. For negative log-likelihood objectives with exponential-family output distributions, this correspondence becomes especially transparent, since the Fisher and Gauss-Newton matrices coincide. We further show that the same local geometric perspective extends to input-space perturbations, where divergence-based regularization is defined through transformations of the input. In this setting, the regularizer induces a pullback quadratic form on the input space, providing a more general perturbation framework than standard SAM while preserving the same local sensitivity interpretation. To validate the analysis empirically, we use the asymmetric $\alpha$-skew Jensen-Shannon divergence (JSD) family as a controlled testbed. Its local curvature coefficient scales as $\alpha(1-\alpha)$ and is maximized at the symmetric point $\alpha=\tfrac12$, which recovers the standard JSD. Loss-landscape visualizations in the input-perturbation regime show that stronger induced curvature penalization is associated with flatter local minima. Experiments on four benchmark datasets further demonstrate that both accuracy and negative log-likelihood are consistently best near this regime of maximal curvature penalization.

What does popular content look like when platforms withhold the usual cues? On TikTok, users still form impressions about which videos are taking off even when likes and view counts are hidden, delayed, or pushed to the margins of the interface. We study this problem through TokOrNot, a web-based game in which participants compared pairs of TikTok videos and reported (i) which one they preferred and (ii) which one they believed had reached a larger audience. We benchmark these judgments against verified public view counts, which we use as a bounded proxy for realized platform reach. Across 3,513 judgments from 363 participants, participants identified the higher-reach video only modestly above chance (56.75%, 95% CI: 56.01-58.55). Preference aligned with the higher-view video at a similar rate, while preference and prediction matched in 83.48% of trials (95% CI: 83.12-85.95). Performance also varied across content categories. Taken together, these results do not suggest that users can reliably read platform success from content alone. Instead, they point to a looser and more uncertain interpretive process in which reach judgments often track personal taste or other weak heuristics when explicit popularity cues are absent. We discuss the implications for algorithmic literacy and for interface designs that reduce visible metrics without leaving users to infer reach from uneven or idiosyncratic cues alone.

Large language models (LLMs) are prone to hallucinating or misinterpreting facts, which impairs their usability in retrieval-augmented generation or data-to-text systems. We analyse how faithfulness of LLMs to provided context depends on how plausible they perceive the context to be (context-memory conflict). To better identify error patterns, we make use of the increased difficulty of non-English and low-resource language text generation and input data based on local knowledge, only partially captured in models' parametric knowledge. We let the models generate text in English, Czech, Slovak and Upper Sorbian from factual (FA), counterfactual (CFA) and fictional (FI) RDF triples containing local Czech and Slovak data. Contrary to our expectations, we observe only a weak context-memory conflict on the human-annotated sample. For Kimi K3 as an LLM judge, which agrees well with human annotations on the sample, counterfactual inputs receive only slightly lower faithfulness scores than factual ones (-0.05 on a 1-5 scale). We also find that a suboptimal choice of LLM judge would lead to overestimating the strength of the context-memory conflict.