Search arXivSearch

SEARCH · Search arXiv

Results for “cs.OS”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

10,400 records · Page 9Linked to original sources

Runtime Safety Filtering for Learned Small UAS Separation Policies under GNSS Degradation

Learning-based separation assurance for small Unmanned Aircraft Systems (sUAS) achieves near-zero collision rates in simulation, but assumes accurate position and velocity information from Global Navigation Satellite Systems (GNSS). This assumption fails in urban environments, where multipath propagation, signal blockage, and intentional interference degrade navigation integrity. This raises a fundamental architectural question for deploying learned separation policies under GNSS degradation: should runtime safety mechanisms filter the policy's actions or its observations? This work evaluates both approaches for multi-agent sUAS separation under adversarial GNSS degradation. Both architectures first estimate a worst-case traffic state consistent with bounded observation uncertainty, then diverge: action filtering constrains policy outputs via discrete-time control barrier functions evaluated at the worst-case state, while observation filtering presents the worst-case state directly to the policy as corrected input. Experimental results show that action filtering provides negligible safety improvement, while observation filtering reduces near mid-air collisions by 90% and remains robust to the barrier function's tradeoff between separation distance and closing rate. These results suggest that, for policies with learned safety behaviors, preserving the policy's decision authority outperforms overriding its actions with hand-designed constraints.

cs.RO

CORE: Improving Compositional Reasoning in MLLM Embedding via Reranker Distillation

MLLM-based embedding models remain limited in compositional retrieval, often failing to distinguish scenes containing the same concepts but different attribute-object bindings. Yet the same backbone can resolve such distinctions when used as a cross-attentive reranker, motivating us to distill its compositional judgments into the embedding model. We propose CORE, which synthesizes candidate lists spanning five compositional matching levels and introduces a Rank-KL objective that trains the embedding model to reproduce the reranker's fine-grained ranking. We further introduce a graded evaluation protocol and compare contrastive learning, pairwise CoSENT, and listwise Rank-KL under the same data and tuning budget. Our comparison shows that both CoSENT and Rank-KL use the multi-level supervision more effectively than contrastive learning, with Rank-KL achieving the strongest overall performance. Across three compositional reasoning benchmarks (COLA, SUGARCREPE++, NEGBENCH), CORE-RERANKER-8B achieves an 82.7% total average, outperforming Jina-Reranker by 10.7 points, while CORE-EMBED-8B achieves the best total average (0.666) among all evaluated embedding models. The improvements transfer to the MCMR benchmark without sacrificing retrieval performance on COCO and Flickr30K.

cs.CV

Mapping of Land Use and Land Cover (LULC) using EuroSAT and Transfer Learning

As the global population continues to expand, the demand for natural resources increases. Unfortunately, human activities account for 23% of greenhouse gas emissions. On a positive note, remote sensing technologies have emerged as a valuable tool in managing our environment. These technologies allow us to monitor land use, plan urban areas, and drive advancements in areas such as agriculture, climate change mitigation, disaster recovery, and environmental monitoring. Recent advances in AI, computer vision, and earth observation data have enabled unprecedented accuracy in land use mapping. By using transfer learning and fine-tuning with RGB bands, we achieved an impressive 99.19% accuracy in land use analysis. Such findings can be used to inform conservation and urban planning policies.

cs.CV

Symbolic Informalization: Fluent, Productive, Multilingual

Symbolic informalization enables a reliable conversion of formal mathematics to natural language. It has the potential to make machine-checked content human-readable without loss of precision. In a traditional proof system usage, symbolic informalization generalizes the limited mechanisms of syntactic sugar into the ordinary language of mathematics. In a setting where proofs are constructed by artificial intelligence and autoformalization, symbolic informalization can explain what precisely has been constructed. This paper outlines the project Informath, which aims to show how symbolic informalization can produce fluent text with a reasonable development effort and address multiple formal and natural languages. Informath is based on an interlingual architecture, where Dedukti works as a hub between different proof systems (Agda, Lean, Rocq) and Grammatical Framework (GF) takes care of linguistic correctness and variation in different natural languages.

cs.AI

SingProbe Technical Report

Runtime guardrails are essential for reliable large language model (LLM) deployment, yet existing approaches typically rely on independent, external models that introduce additional inference cost, delayed safety signals, and a capacity mismatch with increasingly capable base models. To address these issues, we introduce SingProbe, a lightweight intrinsic runtime guard that directly reuses hidden states produced during LLM inference and operates alongside autoregressive decoding. Within a unified framework, SingProbe continuously predicts query intent, response safety, and hallucination risk at the token level with negligible additional guardrail inference overhead, offering a "free-lunch" solution. We further introduce SingStreamBench, a benchmark designed to assess whether streaming guardrails remain inactive on benign prefixes while promptly detecting emerging unsafe content. Extensive experiments show that SingProbe achieves competitive or superior performance compared with substantially larger standalone guardrails and specialized hallucination detectors, with only $\approx$2M parameters and $<0.5\%$ extra overhead. Beyond passive detection, we also show that SingProbe scores can anticipate future generation risk and guide constrained safe decoding. We further extend this paradigm to medical generation through SingProbe-Med, which selectively activates risk-directed decoding interventions only when clinically relevant risks emerge. Together, these results demonstrate that internal model representations provide an effective and efficient interface for generation-time monitoring and control.

cs.CR

LongDS-Bench: On the Failure of Long-Horizon Agentic Data Analysis

Real-world data analysis is inherently iterative, yet existing benchmarks mostly evaluate isolated or short interactive tasks, leaving agents' ability to track evolving analytical context over long horizons untested. We introduce LongDS, a benchmark for long-horizon, multi-turn data analysis where agents must maintain, update, restore, and compose evolving analytical states. LongDS comprises 68 tasks constructed from real-world Kaggle notebooks, spanning 2,225 turns across six domains including Geoscience, Business, and Education. Tasks are designed around state-evolution patterns (e.g., counterfactual perturbation, rollback, multi-state composition), with an average dependency span of 11.3 turns. Evaluating five state-of-the-art models, we find that the best model reaches only 48.45% average accuracy, performance drops nearly 47 points from early to late turns, and long-horizon errors account for 52%--69% of failures. Further analysis shows that additional agent steps do not necessarily improve performance, suggesting that the key bottleneck is maintaining a correct analytical state rather than increasing interaction budget. We release LongDS to support research on reliable long-horizon agentic data analysis. Code and data are released at https://github.com/zjunlp/DataMind.

cs.LG

Predicting Program Exit Code with LLMs and Programming Language Semantics

Large language models (LLMs) have shown proficiency in various software engineering tasks, such as code generation and translation. However, a key limitation in their performance may be their (lack of) understanding of programming-language semantics. Even when explicit semantics are given, it remains unclear whether LLMs apply those rules or lean on priors learned during pre-training instead. We study if LLMs lean on priors or given semantics with a novel task--Program Executability Prediction (PrEx)--that asks models to predict whether a program is semantically valid or invalid (and, if invalid, which formal rule it violates) given the program's syntax and operational semantics. Because PrEx requires both valid and invalid programs, we build a dataset with systematically generated invalid transformations derived from valid programs. We evaluate open-source coding LLMs under two semantic formalisms and two semantic shifts across Human-Written, LLM-Translated, and Fuzzer-Generated program splits. Our findings show that LLMs lean on pre-training priors rather than systematically applying the given rules, performing especially poorly on modified semantics and degrading further as program complexity increases. PrEx is available at https://github.com/EngineeringSoftware/prex.

cs.PL

Hardware-Accelerated Instance Segmentation for Resource-Constrained Space Robotics with Criticality Analysis

Autonomous lunar missions require real-time per- ception under three coupled constraints: extreme low-light conditions, limited onboard compute, and radiation-induced hardware faults that can silently corrupt inference. We present a deployment-oriented instance segmentation framework for resource-constrained lunar robotics that jointly addresses quan- tization calibration and system-level fault exposure under strict compute constraints. First, we introduce Activation Variance Informative Sampling (AVIS), a label-free calibration strategy that deterministically selects calibration samples based on activation variance statistics. Second, we deploy a YOLO-based segmentation model on a Deep Learning Processor Unit (DPU) with architectural modifications that reduce CPU fallback paths and enable statically compiled execution with bounded latency in low-lighting conditions. We further introduce a software-level criticality analysis to estimate fault exposure and guide mitigation under radiation-constrained operation. On a lunar micro-rover platform, AVIS with bias correction recovers 69.8% of quantization-induced accuracy loss while achieving 309 ms inference latency and 5.7 W power consumption. Targeted mitigation reduces global criticality by 31.7%. The results demonstrate an integrated approach and a blueprint for a reliable and safe AI perception framework under space deployment constraints.

cs.RO

TUX: Measuring Human--AI Tacit Understanding

As large language models (LLMs) increasingly act as collaborative partners, human--AI alignment is often evaluated through explicit task success, accuracy, or reward optimization. Yet many collaborative settings depend on tacit understanding: whether an agent can align with a human's evaluative stance or representational priors without clear objectives, communication, or feedback. To study this capacity, we develop a spectrum-placement task inspired by the social party game Wavelength, in which humans and agents independently place concepts along subjective spectra. We operationalize the Tacit Understanding Index (TUX) as a pairwise behavioral measure of similarity between human and agent judgments, and evaluate it with 241 human participants and 200 profile-conditioned LLM agents across four models. We find that nearest human--agent pairs in trait space achieve significantly higher TUX, suggesting that tacit alignment is associated with person-level characteristics rather than reflecting only random similarity. Regression analyses show that TUX becomes more explainable as predictor sets become richer, with individual traits, decision-making styles, and confidence improving over aggregate trait-distance baselines. These findings suggest that TUX provides a measurable behavioral signal of human--LLM tacit understanding, while revealing the limits of profile-based conditioning for capturing deeper representational alignment.

cs.HC

GazeRefine: Expert Gaze as a Test-Time Prompt for Training-Free Medical Image Segmentation

Medical image segmentation remains difficult to scale because high-performing methods typically rely on dense expert annotations and task-specific training. We introduce GazeRefine, a training-free framework that uses gaze as an inference-time prompt for zero-shot medical image segmentation. Sparse, duration-weighted fixations are converted into foreground and background priors that initialize semantic prototypes in frozen DINOv3 feature space. These prototypes are iteratively refined through foreground-background discrimination, feature-space affinity propagation, and anchoring to the initial gaze guidance, allowing segmentation to extend beyond directly fixated regions while limiting semantic drift. GazeRefine requires no segmentation masks, fine-tuning, adapters, prompt encoders, or gradient updates. We evaluate the method on gaze-annotated polyp segmentation and prostate MRI segmentation. The results show strong performance on colonoscopy images and competitive performance on prostate MRI, supporting gaze-guided prototype refinement as a promising approach for segmentation-label-efficient, human-in-the-loop medical image segmentation. Our tools and code can be found in the following repository: https://github.com/MohammedOussamaBEN/GazeRefine.git

eess.IV

Edge Geography is XNLP-hard for Pathwidth and in XP for Tree-Partition Width

Directed Edge Geography and Undirected Edge Geography are classical PSPACE-complete two-player graph games in which players alternately make moves along edges, deleting each one after use; the first player unable to move loses. We prove that both problems are XNLP-hard when parameterized by pathwidth, addressing a question raised by Bodlaender over 30 years ago. On the positive side, we observe that Directed Edge Geography is fixed-parameter tractable when parameterized by treewidth and maximum degree. We also prove that both problems are in XP on simple graphs when parameterized by tree-partition width. These results develop modern lower-bound and decomposition-based algorithmic methods for width-based questions in PSPACE-complete graph games.

cs.CC

Mapping the Reddit Bot Ecosystem: Taxonomy and Evolution

Automated agents increasingly participate in online communities, yet their population structure and roles remain poorly understood. Using a dataset of 3,389 identified bots and their full activity histories, we construct a taxonomy of bot "species" on the news aggregation and social media platform Reddit based on temporal, community, linguistic, and semantic features. Clustering analysis reveals 18 distinct bot types spanning content-specialized, behavior-driven, and infrastructural roles such as moderation and utility support. In addition, temporal analysis shows that bot numbers and activity expanded rapidly before peaking around the COVID-19 period, then started declining even before Reddit's 2023 API policy changes. However, the overall diversity of bot species has remained remarkably stable. These findings suggest that online bot populations form evolving digital ecosystems.

cs.SI

Beyond Helpfulness: A Teaching-over-Solving Diagnostic for Measuring Educational Impact in LLM Tutors

Large language models are increasingly proposed as educational tutors, yet stronger task-solving ability does not necessarily imply stronger learning support. Motivated by recent calls to measure the social impact of NLP systems in practice, we study whether public LLM tutoring benchmarks distinguish learning-supportive behavior from mere answer production. We propose a lightweight diagnostic based on the gap between solving-oriented and pedagogy-oriented benchmark performance. Using public MathTutorBench leaderboard results, we show that these dimensions are only partially aligned: across eight publicly reported models, the correlation between solving and pedagogy composites is 0.421, and several models shift meaningfully in rank when evaluation moves from solving to pedagogy. We then analyze the public TutorBench sample and show that agency-relevant behaviors are explicitly encoded in benchmark rubrics, especially in active-learning settings that reward guiding questions, calibrated hints, and non-disclosive scaffolding. Together, these findings suggest that educational-impact evaluation should not treat task success as a sufficient proxy for learning support. We argue that public tutoring benchmarks can better support positive-impact evaluation by reporting solving-oriented and pedagogy-oriented scores separately and by making disclosure-sensitive, student-agency-preserving criteria more explicit.

cs.AI

Large-Scale User Behavior Analysis in Multimodal AI-Assisted Manual Task Execution

Conversational Task Assistants (CTAs) are multimodal dialogue systems that support users in complex real-world tasks such as cooking and DIY through voice, text, image, and video interactions. Prior user studies have focused on controlled settings, leaving limited understanding of real-world CTA usage at scale. In this work, we present a large-scale study of CTA usage based on thousands of users in-the-wild. Our large-scale real-world data analysis unveils new understandings of (i) user-CTA interaction flows, (ii) user intents, (iii) user conversational traits, and (iv) behavioral factors associated with user satisfaction. Our findings reveal key opportunities for future research in CTAs, particularly in user interaction design and task engagement, concluding with concrete design guidelines.

cs.HC

Neural Symbollic Regression Using Deep Learning and Sparse Modelling

Symbolic Regression (SR) seeks to find succinct mathematical expressions that represent the fundamental relationships within data, providing interpretability and scientific understanding that exceeds that of black-box models. Nevertheless, traditional methods like Genetic Programming face challenges with scalability and are highly sensitive to noise, while sparse regression techniques such as SINDy rely significantly on predetermined feature libraries. In this work, we present a Neural Symbolic Regression (NSR) framework that treats neural networks as functional preconditioners for symbolic discovery. Our approach uses a decoupled pipeline: a neural network first learns a smooth, noise-robust approximation of the target function in an interaction- aware nonlinear feature space. LASSO is then applied to extract sparse, interpretable closed-form expressions. To improve predictive accuracy and symbolic fidelity by integrating distributed hyperparameter optimization with Ray Tune and ASHA scheduling. Experiments on the Nguyen benchmark suite show that our approach consistently outperforms SINDy and non-tuned neural baselines in RMSE, noise robustness, and out-of-distribution generalization. Ablation studies confirm the significance of feature interactions, neural depth, and tuning strategies. In general, this study presents a scalable and understandable neural-symbolic framework, creating a solid link between neural approximation and the discovery of sparse equations for scientific machine learning.

cs.LG

Scalable FPGA Framework for Real-Time Denoising in High-Throughput Imaging: A DRAM-Optimized Pipeline using High-Level Synthesis

High-throughput imaging workflows, such as Parallel Rapid Imaging with Spectroscopic Mapping (PRISM), generate data at rates that exceed conventional real-time processing capabilities. We present a scalable FPGA-based preprocessing pipeline for real-time denoising, implemented via High-Level Synthesis (HLS) and optimized for DRAM-backed buffering. Our architecture performs frame subtraction and averaging directly on streamed image data, minimizing latency through burst-mode AXI4 interfaces. The resulting kernel operates below the inter-frame interval, enabling inline denoising and reducing dataset size for downstream CPU/GPU analysis. Validated under PRISM-scale acquisition, this modular FPGA framework offers a practical solution for latency-sensitive imaging workflows in spectroscopy and microscopy.

cs.AR

ViTAMINS: An Empirical Study of Training Self-Supervised Vision Transformers with Synthetic Hard Negatives

We introduce ViTAMINS, a method that integrates synthetic hard negatives into unsupervised vision transformer pretraining to improve representation quality. Our approach is thoroughly benchmarked on ImageNet and transfer learning, image retrieval, copy detection, and image, video segmentation tasks. Notably, our proposed negatives give rise to emergent properties, where learned representations contain explicit information about the semantic content of an image and serve as excellent classifiers (up to +11.3% over baselines). ViTAMINS achieves these benefits through simple modifications to existing contrastive frameworks and outperforms competing methods while being more resource efficient, e.g., our ViT-B surpasses V-JEPA with ViT-L. Our findings motivate reconsidering contrastive learning as a simpler yet powerful alternative to dominant generative and self-distillation approaches.

cs.CV