AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchBR-MoE: Budgeted Semantic Release for Privacy-Preserving Traffic Analytics
arXiv:2610.27026v2 Announce Type: replace-cross Abstract: Traffic cameras support useful analysis, but their outputs can reveal private information. Replacing images with text reduces direct visual exposure. It does not…
- arXiv cs.AIresearchRobustness of Internal Signals for Hallucination Detection in Vision-Language Models
arXiv:2610.27041v1 Announce Type: cross Abstract: When given an image and a text prompt, vision-language models (VLMs) can produce seemingly sophisticated text responses. However, they are also prone to hallucinate objec…
- arXiv cs.AIresearchCross-Lingual Reusability of LLM Pipelines: Financial Extraction, Multilingual Reasoning, and Semantic Screening
arXiv:2610.27043v1 Announce Type: cross Abstract: Financial institutions increasingly deploy large language models (LLMs) in multilingual markets, raising a practical question: can pipelines developed and validated in En…
- arXiv cs.AIresearchDiscrete Neural Interlingua for Ontology-Agnostic EHR Modeling
arXiv:2610.27047v1 Announce Type: cross Abstract: EHR foundation models promise scalable clinical prediction across health systems, yet their deployment is constrained by a tokenization bottleneck. The same clinical conc…
- arXiv cs.AIresearchJoint Constraint Aggregation for Qualifier-Sensitive Retrieval-Augmented QA
arXiv:2610.27114v1 Announce Type: cross Abstract: Retrieval-augmented QA can fail even when the correct answer is in the retrieved evidence. This can happen when a competing answer satisfies some, but not all, of the que…
- arXiv cs.AIresearchReconstruction-Optimized Expert Pruning for Mixture-of-Experts Language Models
arXiv:2610.27123v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) models reduce computation through sparse expert activation, but deployment still requires storing a large number of expert parameters. We study p…
- arXiv cs.AIresearchMind Your Boundaries: Boundary-Aware Model Obfuscation against Cross-node Attack
arXiv:2610.27135v1 Announce Type: cross Abstract: With the advancement of powerful AI chips, on-device machine learning is gaining traction for its low latency and privacy protection. However, when edge devices are compr…
- arXiv cs.AIresearchOntic state recovery through prediction
arXiv:2610.27139v1 Announce Type: cross Abstract: We encounter a system through its responses, yet its state concerns how the system is. Predicting further responses promises knowledge beyond what is immediately observed…
- arXiv cs.AIresearchRepresenting short-term movement to reveal animal identity and learning progress
arXiv:2610.27140v1 Announce Type: cross Abstract: How learning progresses over time and varies across individuals is a longstanding question in cognitive and behavioral science. However, a major challenge is identifying…
- arXiv cs.AIresearchICLHerd: Herding-Based Data Selection for LLMs with In-Context Learning as a Proxy
arXiv:2610.27142v1 Announce Type: cross Abstract: Data selection plays a central role in modern large language model training, as the quality of training data strongly influences downstream performance. A common approach…
- arXiv cs.AIresearchModel Exposure on One Device Endangers Others: Rethinking Edge Neural Network Security across Devices
arXiv:2610.27152v1 Announce Type: cross Abstract: Neural networks (NNs) are increasingly deployed on edge devices such as mobile and IoT platforms, offering low latency and improved privacy compared to cloud-based deploy…
- arXiv cs.AIresearchUnderstanding the Benefits of Scaffolds through Computational Separations
arXiv:2610.27157v1 Announce Type: cross Abstract: Inference-time harnesses, i.e., external procedures that call a model many times and keep state across calls, can substantially improve model capabilities. In this paper,…
- arXiv cs.AIresearchOn the Fragility of Tasks under Composition: Implicit Budgeting in Language Models
arXiv:2610.27190v1 Announce Type: cross Abstract: Modern language models are increasingly asked to handle several tasks within a single prompt. Doing so requires solving each task correctly while integrating the resultin…
- arXiv cs.AIresearchCARLAR Influence: A Structured Hallucination Risk in Multimodal Cheque Extraction and an Exploratory Compressed-Image Countermeasure
arXiv:2610.27192v1 Announce Type: cross Abstract: Multimodal large language models (LLMs) are increasingly used for structured extraction from financial documents, but their behavior under visual ambiguity is poorly unde…
- arXiv cs.AIresearchLanguage Guidance for Supervised Vision Training: An Empirical Study of Generalization
arXiv:2610.27208v1 Announce Type: cross Abstract: Deep neural networks have achieved remarkable success on vision benchmarks, yet they continue to struggle with many generalization challenges. Supervised vision training…
- arXiv cs.AIresearchProfile-DP: Differential Privacy for User-Controlled Profile Release in Personalized LLM Inference
arXiv:2610.27213v1 Announce Type: cross Abstract: Personalized large language models (LLMs) increasingly rely on profile-conditioned inference, pairing public requests with rich, user-specific facts. This paradigm expose…
- arXiv cs.AIresearchTarget-Aware Curation of Hallucination Stress Tests for Large Language Models
arXiv:2610.27247v1 Announce Type: cross Abstract: Hallucination stress tests can yield similar average scores across language models even when different prompts elicit hallucinations in each model. We develop a procedure…
- arXiv cs.AIresearchValiAssist: Scaling LLM-generated Programming Feedback Validation by Demonstration
arXiv:2610.27248v1 Announce Type: cross Abstract: Generative AI can create programming feedback at scale, but validating its alignment with instructors' pedagogical goals and strategies remains challenging. Instructors m…
- arXiv cs.AIresearchWhen Models Know but Can't Linearly Say: Relational Encoding Limits Linear Vulnerability Classification in Code LLMs
arXiv:2610.27257v1 Announce Type: cross Abstract: Large language models can detect vulnerable code, yet the form of vulnerability encoding in their representations remains unclear. We analyze 8 code models spanning 6.7B-…
- arXiv cs.AIresearchTheorem prover Arend
arXiv:2610.27264v2 Announce Type: replace-cross Abstract: This paper is an exposition of the theorem prover Arend developed at JetBrains Research. Arend language has been designed to facilitate the formalization of mathe…
- arXiv cs.AIresearchAuditing Safety Preference Supervision with Recorded Derivations
arXiv:2610.27266v1 Announce Type: cross Abstract: Preference labels align model training, but the judgments behind them are difficult to inspect or verify. In this paper, we examine how preference construction can be che…
- arXiv cs.AIresearchVPFlow: Value-Predictive Planning with Imperfect World Models for Offline Goal Reaching
arXiv:2610.27280v1 Announce Type: cross Abstract: Offline goal-conditioned reinforcement learning enables skill acquisition by reaching any goal from any state without dense rewards. Model-based methods allow for plannin…
- arXiv cs.AIresearchWhen Does Hate Transfer? A Large-Scale Study of Cross-Lingual Hate Speech Detection
arXiv:2610.27285v1 Announce Type: cross Abstract: Effective detection of hate content is essential for maintaining healthy online communities, yet performance can vary substantially across languages. This work investigat…
- arXiv cs.AIresearchGeneralizing Multimodal Foundational Models in Context-Driven Class Incremental Learning
arXiv:2610.27298v1 Announce Type: cross Abstract: Class-Incremental Learning (CIL) aims to learn new classes continuously without forgetting the previously learned ones. Recent works in CIL have significantly advanced ac…
- arXiv cs.AIresearchGlobal Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages
arXiv:2610.27304v1 Announce Type: cross Abstract: Large language models are deployed globally, yet safety alignment and evaluation remain concentrated in English: harmful requests refused in English can elicit unsafe com…
- arXiv cs.AIresearchSteering Vision-Language Models for Visual-Contextual Privacy in Location Disclosure
arXiv:2610.27323v1 Announce Type: cross Abstract: Vision-language models (VLMs) can infer sensitive information such as geolocation from subtle visual cues, creating privacy risks when responses disclose more detail than…
- arXiv cs.AIresearchExposing Vulnerability Monoculture Through Systemic Adversarial Robustness
arXiv:2610.27353v1 Announce Type: cross Abstract: Beneath the apparent diversity of foundation models lies an increasingly homogeneous technical stack of common training pipelines, distillation methods, and data. These s…
- arXiv cs.AIresearchLiBRA: Logit Fingerprints for Byzantine-Robust Client Selection in Federated Learning
arXiv:2610.27357v1 Announce Type: cross Abstract: Byzantine-robust aggregation shields federated learning from malicious updates, but the rules that filter out suspect updates do so by comparing clients' parameter vector…
- arXiv cs.AIresearchOn the Limits of Scale: Inverse Scaling in Vision-Language Models
arXiv:2610.27360v1 Announce Type: cross Abstract: Scaling laws suggest that increasing model size, training data, and compute generally yields monotonic performance improvements. While this often holds at the aggregate l…
- arXiv cs.AIresearchPredictive Closure: A Set-Level Characterization of JEPA's Feature Selection
arXiv:2610.27364v1 Announce Type: cross Abstract: In joint-embedding predictive architectures (JEPA), an encoder and a predictor are trained together to produce a representation that minimizes the error in predicting the…
- arXiv cs.AIresearchTrail-Weighted Influence Scores for Approximate Outcome Prediction in Bayesian Networks
arXiv:2610.27366v1 Announce Type: cross Abstract: We study trail-weighted pruning of evidence for Bayesian-network prediction. A traversal enumerates the active trails from each evidence node to the outcome, scores each…
- arXiv cs.AIresearchFundamental Impossibility Results in AI Safety: Trading Off between Desiderata
arXiv:2610.27373v1 Announce Type: cross Abstract: Is there a fundamental tradeoff between desiderata for safe AI systems, such as harmlessness and helpfulness? We answer by formalizing the interaction between AI systems…
- arXiv cs.AIresearchThe Router Already Knows
arXiv:2610.27421v1 Announce Type: cross Abstract: Mixture-of-experts (MoE) models already make a sparse choice for every token: which experts should process it. We ask whether a sparse dictionary should reuse that choice…
- arXiv cs.AIresearchMLIPS: Multi-Layer Indexer Pool Sharing
arXiv:2610.27438v2 Announce Type: replace-cross Abstract: Sparse attention limits the keys used by each query, but finding those keys can still require a full-prefix scan at every layer. We introduce $MLIPS:$ $Multi$-$La…
- arXiv cs.AIresearchScaling Graph Reasoning in Language Models with Learnable Graph Positional Encodings
arXiv:2610.27439v1 Announce Type: cross Abstract: Solving reasoning tasks supported on graphs typically requires providing a Language Reasoning Model (LRM) with a serialized description of the graph. This approach result…
- arXiv cs.AIresearchBeyond Scalar Leaderboards: Adaptive Sampling for Pareto Frontier Identification in Multi-Objective LLM Arenas
arXiv:2610.27447v1 Announce Type: cross Abstract: The evaluation of large language models (LLMs) is inherently multi-objective, as different models may offer different tradeoffs across performance dimensions such as math…
- arXiv cs.AIresearchProjected Distillation: On Learning to Solve Hard Problems from Close, Correct Targets
arXiv:2610.27467v1 Announce Type: cross Abstract: On-policy learning is central to post-training: training on a model's own generations limits distribution shift and helps preserve existing capabilities. On hard problems…
- arXiv cs.AIresearchTraining-Free Continual Learning for Multimodal Forecasting via Prompt Evolution
arXiv:2610.27468v1 Announce Type: cross Abstract: Adapting a forecasting system to a changing world classically means retraining. We ask whether the task of continually learning regimes for time series forecasting can be…
- arXiv cs.AIresearchPrivacy-Preserving Personalized Federated Learning (PPP-FL) for Medical Imaging
arXiv:2610.27470v1 Announce Type: cross Abstract: Federated learning (FL) is a distributed machine learning paradigm that reduces transmission costs, mitigates cross-border data sharing restrictions, and enhances privacy…
- arXiv cs.AIresearchBeyond the Prompt: Recovering Hidden Concepts in Text-to-Image Models
arXiv:2610.27473v1 Announce Type: cross Abstract: Text-to-image models can generate visual concepts that their prompts never mention, and recent work shows that such associations can be implanted so that ordinary prompts…
- arXiv cs.AIresearchOrthogonal Boundary Unlearning for Dual-Use Concepts in Large Language Models
arXiv:2610.27490v1 Announce Type: cross Abstract: Mitigating catastrophic risks from large language models requires excising hazardous operational capabilities, such as cyberwarfare exploitation and pathogen enhancement,…
- arXiv cs.AIresearchCross-Visibility-Gap Identity Persistence for 4D Scene Graphs
arXiv:2610.27505v1 Announce Type: cross Abstract: 4D scene graphs require persistent object identity when an object disappears inside an opaque carrier and later reappears as a fresh detection. When appearance and target…
- arXiv cs.AIresearchProgressive Subset Selection and Cascaded Consistency for Unsupervised Vision-Language Model Adaptation
arXiv:2610.27515v1 Announce Type: cross Abstract: Pretrained Vision-Language Models (PVLMs) achieve strong zero-shot recognition but often suffer performance degradation under target-domain shift. Unsupervised adaptation…
- arXiv cs.AIresearchHealBench: A Fine-Grained Benchmark for Evaluating Psychological Comfort in Text-to-Image Generation
arXiv:2610.27516v1 Announce Type: cross Abstract: Despite advances in photorealism, semantic fidelity, and aesthetics, text-to-image generation and image editing remain understudied in a human-centered dimension: whether…
- arXiv cs.AIresearchAdversarial Benchmarks and BenchDoor Attacks: Exploiting Benchmark Incentives for Backdoor Attacks
arXiv:2610.27533v1 Announce Type: cross Abstract: Machine learning benchmarks act as guideposts for model development and spur the creation of training datasets intended to improve benchmark scores. We introduce BenchDoo…
- arXiv cs.AIresearchLess is More: Concise Retrieval and Critic-based Re-ranking for Composed Image Retrieval
arXiv:2610.27538v1 Announce Type: cross Abstract: Zero-shot composed image retrieval (ZS-CIR) aims to retrieve a target image based on a reference image and modification text. Recent approaches commonly rely on large vis…
- arXiv cs.AIresearchCan Native Safety Travel? Reusing Safety Pretraining Across Models via On-Policy Distillation
arXiv:2610.27555v1 Announce Type: cross Abstract: Safety pretraining has emerged as a promising paradigm for embedding safety deeply into a model during pretraining, yielding robustness to subsequent fine-tuning that con…
- arXiv cs.AIresearchAnchorRM: Parallel Intra-Batch State Reconstruction with Anchor-Based Residual Memory for Temporal Graph Neural Networks
arXiv:2610.27558v1 Announce Type: cross Abstract: Memory-based temporal graph neural networks (TGN) rely on recurrent node memories to capture evolving interaction histories. However, temporal mini-batching introduces a…
- arXiv cs.AIresearchCityPaint: A Free Lunch from 2D Diffusion for City-Scale 3D Texturing
arXiv:2610.27564v1 Announce Type: cross Abstract: Generating high-fidelity, city-scale 3D environments plays a crucial role in applications such as virtual reality, gaming, and urban digital twins. However, the scarcity…
- arXiv cs.AIresearchAnomalyJev: Towards Evaluating Image Anomaly Explanations via Claim-Level Verification
arXiv:2610.27565v1 Announce Type: cross Abstract: Recent multimodal LLM (MLLM)-based anomaly detection (AD) methods can generate textual explanations that help interpret their predictions. However, reliably and efficient…