AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchLLaCO: Continuous Diffusion for Large Language Models with Competition-Aware Semantic Scheduling
arXiv:2610.32604v1 Announce Type: new Abstract: Continuous diffusion models have emerged as a promising non-autoregressive paradigm for text generation over continuous embedding spaces. Along the continuous trajectory, t…
- arXiv cs.AIresearchLGMT: A Tool for Testing the Logical Reasoning Reliability of LLMs
arXiv:2610.32627v1 Announce Type: new Abstract: Large Language Models (LLMs) increasingly serve as reasoning components in modern software systems, yet a model that answers a logical query correctly can contradict itself…
- arXiv cs.AIresearchWiMoT: Replay-Free Continual Wi-Fi Sensing via Moment Transport
arXiv:2610.32660v2 Announce Type: replace Abstract: Continual Wi-Fi sensing must balance plasticity, learning new rooms or users, against stability, keeping earlier ones, yet nodes often cannot retain channel state infor…
- arXiv cs.AIresearchDo Language Models Trust Their Own Justifications? A Study on Functional Consistency
arXiv:2610.32665v1 Announce Type: new Abstract: Large Language Models (LLMs) have been widely adopted in text classification tasks, where they not only output class predictions but also generate explanations that highlig…
- arXiv cs.AIresearchRetiming Robot Demonstrations with a Learned Speed Prior
arXiv:2610.32676v1 Announce Type: new Abstract: The pace of a robot demonstration is often determined by the teleoperator rather than by the task itself. As a result, the same motion can appear at widely different speeds…
- arXiv cs.AIresearchA Foundation Model for Policy In-Context Longitudinal Causal Effect Estimation
arXiv:2610.32863v1 Announce Type: new Abstract: Longitudinal causal effect estimation is complicated by time-varying confounding and treatment-confounder feedback loops. Existing methods generally require data-specific n…
- arXiv cs.AIresearchFrom Deliberately Constructed Systemic Representations of the World of Activity to Dialogue: A Five-Level Communication Architecture for Strong Language-Based ASI
arXiv:2610.32869v1 Announce Type: new Abstract: This paper proposes a five-level organization of message formation and understanding within thinking and activity in which ASI participates. Its central thesis is that a la…
- arXiv cs.AIresearchAdvancing SVD-based LLM Compression via Layer-Wise Error Model Search
arXiv:2610.32878v2 Announce Type: replace Abstract: Low-rank SVD-based compression offers a powerful strategy to reduce the computational costs of LLMs. However, existing methods face two key limitations: (i) global rank…
- arXiv cs.AIresearchNgramRL: Accelerating RL Post-Training for Structured Generation via Rollout-Aware N-gram Speculative Decoding
arXiv:2610.32881v1 Announce Type: new Abstract: Reinforcement learning (RL) post-training improves the reasoning capabilities of large language models, but its cost is dominated by autoregressive rollout generation, whic…
- arXiv cs.AIresearchClosing the Modality Gap via Simplex-Constrained Representations
arXiv:2610.32890v1 Announce Type: new Abstract: While multimodal contrastive models like CLIP and CLAP successfully learn shared representations, their embeddings typically exhibit a persistent \emph{modality gap}. We pr…
- arXiv cs.AIresearchAstro2Method: Benchmarking LLMs on Machine Learning Method Selection in Astronomy
arXiv:2610.30282v1 Announce Type: cross Abstract: Machine learning (ML) is increasingly used in astronomy, but choosing a suitable method for a scientific problem often requires expertise in both astronomy and ML. Large…
- arXiv cs.AIresearchValidating the Comparison, Not Just the Scorer: Why Overall Accuracy Can Mislead in Interactive LLM Evaluation
arXiv:2610.31678v1 Announce Type: cross Abstract: Many LLM studies compare how models respond to different follow-up prompts. For example, one prompt may neutrally ask a model to reconsider, while another insists that an…
- arXiv cs.AIresearchRethinking Structured Generation: Can Graph-Based Reasoning Resolve Ambiguity?
arXiv:2610.31681v1 Announce Type: cross Abstract: Structured generation is typically formulated as a deterministic mapping from input to a single structured output. This assumption is fundamentally misaligned with natura…
- arXiv cs.AIresearchFrom Target Selection to Condition Reuse: Auditing Residual Interventions in Tool-Use Models
arXiv:2610.31686v1 Announce Type: cross Abstract: Activation interventions can make a language model produce a target answer, but can the model use the intended condition in other computations? We study this question in…
- arXiv cs.AIresearchDistributional Alignment of LLMs and Its Generalization Properties
arXiv:2610.31687v1 Announce Type: cross Abstract: Large language models are increasingly used to simulate and probe human behavior, e.g., for surveys. For these applications, a key property is that outputs satisfy distri…
- arXiv cs.AIresearchCollaborative In-Context Learning for Indoor Radio Map Estimation
arXiv:2610.31702v1 Announce Type: cross Abstract: Indoor radio map estimation (RME) is essential for wireless applications such as coverage planning, localization, and network deployment. However, most existing methods f…
- arXiv cs.AIresearchEnd-to-End Differentiable Diffusion Conditioning for Physics-Informed Optimization
arXiv:2610.31721v1 Announce Type: cross Abstract: Optimizing high-dimensional physical systems under expensive evaluations and complex constraints is a central challenge in science and engineering. In offline model-based…
- arXiv cs.AIresearchAutoregressive Multimodal Anomaly Detection with Large Language Models for Microservice Systems
arXiv:2610.31722v1 Announce Type: cross Abstract: Microservice systems have become a dominant architecture for large-scale cloud applications, enabling rapid deployment and elastic scaling. However, ensuring the reliabil…
- arXiv cs.AIresearchLearning to Be Monitored: Causal Bounds and Intervention-Aware Post-Training Recipes for Chain-of-Thought Monitorability
arXiv:2610.31733v1 Announce Type: cross Abstract: Chain-of-thought (CoT) monitoring aims to detect potentially harmful behaviors in large language models (LLMs) from their CoTs. Intervention-based evaluations measure thi…
- arXiv cs.AIresearchGeneralized green and red concepts for artificial intelligence, big data, and information and communication technologies: Multidimensional framework and countermeasures
arXiv:2610.31739v1 Announce Type: cross Abstract: The rapid expansions of artificial intelligence (AI), big data (BD), and information and communication technologies (ICT) are driving substantial growth in electricity, w…
- arXiv cs.AIresearchPrefO2O: Zero-Shot Preference-Based Reinforcement Learning via Offline-to-Online Latent Preference Transfer
arXiv:2610.31812v1 Announce Type: cross Abstract: Preference-based reinforcement learning (PbRL) has emerged as a promising paradigm for learning robot behaviors aligned with human intent. However, its scalability is fun…
- arXiv cs.AIresearchWhat the Query Remembers: Query Abstraction Governs Lexical and Dense Retrieval of Classical Arabic Poetry
arXiv:2610.31821v1 Announce Type: cross Abstract: How well a poetry archive can be searched depends on what the user still remembers, not only on what the model understands. We show this for Classical Arabic poetry, benc…
- arXiv cs.AIresearchSEVI: Safe Evidential Real-World Reinforcement Learning
arXiv:2610.31829v1 Announce Type: cross Abstract: Real-world reinforcement learning has enabled contact-rich manipulation tasks to be trained directly on physical robots, but unsafe interactions during online learning re…
- arXiv cs.AIresearchRethinking Multi-View Fusion for Urban Representation Learning
arXiv:2610.31850v2 Announce Type: replace-cross Abstract: Urban region representation learning encodes heterogeneous urban signals, such as mobility flows, points of interest, and land-use patterns, into unified embeddin…
- arXiv cs.AIresearchMachine-Readable Operator Manuals over the ISO 11783 Network for Offline, Multilingual LLM Assistance
arXiv:2610.31863v1 Announce Type: cross Abstract: If agricultural machinery is to support intelligent, offline assistance, then the system must access technical documentation dynamically. In this paper, the Semantic Sync…
- arXiv cs.AIresearchFSVL: Taxonomic Fast-Slow Vision-to-Language Collaborative Framework for Interpretable Confusable Fish Species Identification and Knowledge Generation in the Middle Yangtze River
arXiv:2610.31880v1 Announce Type: cross Abstract: The Yangtze River Basin is a critical region for freshwater biodiversity conservation. The implementation of the Yangtze River Ten-Year Fishing Ban has increased the impo…
- arXiv cs.AIresearchEffectiveness of Curriculum Learning Depends on Reward Sparsity and Competing Optima
arXiv:2610.31884v1 Announce Type: cross Abstract: While animals and people tend to learn a task more quickly and reliably when they first train on simpler versions of it, curriculum learning's effectiveness in artificial…
- arXiv cs.AIresearchLearning Structured Sparsity for Large Models: Bridging Model Compressibility and Efficient Inference
arXiv:2610.31897v1 Announce Type: cross Abstract: Scaling laws continue to reward increases in model scale, while the resulting computational and memory demands drive greater reliance on cloud infrastructure for inferenc…
- arXiv cs.AIresearchENCL: Enabling Efficient Neuromorphic Continual Learning with Local Learning Rule and Sub-Network Freezing Strategy
arXiv:2610.31927v1 Announce Type: cross Abstract: Continual learning (CL) in neural networks is a promising learning approach for updating the knowledge of deployed systems based on new data encountered at run time, that…
- arXiv cs.AIresearchKolibri: A Sovereign European Model on the Pareto Frontier
arXiv:2610.31945v1 Announce Type: cross Abstract: This report introduces Kolibri, an English-German mixture-of-experts transformer with 78.1B total parameters and 3.46B active parameters per token, released as open weigh…
- arXiv cs.AIresearchRobustNCL: Robustifying Neuromorphic Continual Learning under Transient Faults with Lightweight Detect-and-Cure Strategy
arXiv:2610.31954v1 Announce Type: cross Abstract: In neural networks, continual learning (CL) is a prominent approach for enabling deployed systems to adapt to dynamic operational environments by effectively learning fro…
- arXiv cs.AIresearchA Mathematical Theory of the Jacobian Lens
arXiv:2610.31955v1 Announce Type: cross Abstract: Recent work from Anthropic introduced a new interpretability tool, called the Jacobian lens (J-lens, Gurnee et al., 2026). This method identified linear directions in the…
- arXiv cs.AIresearchSpectra: Multi-Scale Implicit Decomposition for Deepfake Detection
arXiv:2610.31979v1 Announce Type: cross Abstract: Deepfakes increasingly undermine trust in digital media, yet existing detectors often struggle to generalize across unseen manipulation techniques, compression levels, an…
- arXiv cs.AIresearchOpenHome: From Perception to Insight in Smart-Home Camera Videos
arXiv:2610.31986v1 Announce Type: cross Abstract: Vision language models (VLMs) have been increasingly adopted in consumer smart-home camera systems to interpret recorded videos and provide insights. However, video clips…
- arXiv cs.AIresearchMI-MIDI 2: From Observation to Control with Sparse Autoencoders in Text-to-MIDI Generation
arXiv:2610.32007v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) can expose features of a generative model without predefined labels, but in language models the evidence used to interpret these features has p…
- arXiv cs.AIresearchRethinking Representation Geometry in Supervised Contrastive Learning under Long-Tailed Data
arXiv:2610.32039v1 Announce Type: cross Abstract: Supervised contrastive learning and its balanced variants have demonstrated their effectiveness for representation learning under long-tailed distributions. However, alth…
- arXiv cs.AIresearchACVA-Bench: Beyond Prompt Following in Abstract Concept Visualization Ability for Text-to-Image Models
arXiv:2610.32058v1 Announce Type: cross Abstract: Text-to-image (T2I) models have achieved remarkable progress in image quality and instruction following, prompting growing research interest in their capacity for deep un…
- arXiv cs.AIresearchConcept Acupuncturist: An Efficient and Precise Concept Erasure Method for Diffusion Models via Sparse Autoencoders
arXiv:2610.32059v1 Announce Type: cross Abstract: Diffusion-based image generation models produce high-quality images but may generate harmful or policy-violating content, motivating machine unlearning (MU) to selectivel…
- arXiv cs.AIresearchControlled Evaluation of EEG Foundation Models: Random Initialisation, Band-Power and Validation-Selected Probes
arXiv:2610.32067v1 Announce Type: cross Abstract: EEG foundation models report downstream gains, yet their protocols seldom include three controls. These are the same encoder untrained, spectral features fitted on the sa…
- arXiv cs.AIresearchGuided Zero-Shot Anomaly Explanation for Anonymized Crowd Video Monitoring: Context, Categorization, and Spatial Localization
arXiv:2610.32073v1 Announce Type: cross Abstract: Despite the promising development of deep learning for video anomaly detection (VAD), explainability (XAI) of these models remains a challenge. Explainable and privacy-pr…
- arXiv cs.AIresearchTTT in VLA Models: Exploring Training with Human Demonstrations
arXiv:2610.32094v1 Announce Type: cross Abstract: Long-horizon robot tasks require retaining and using past information. We integrate Test-Time Training (TTT), which stores context in fast weights, into InternVLA-A1 and…
- arXiv cs.AIresearchReMedI: Inference-Time Refinement and Self-Distillation for Controllable Molecular Optimization
arXiv:2610.32103v1 Announce Type: cross Abstract: In drug development, lead optimization must improve competing properties, such as increasing permeability and reducing toxicity, while limiting scaffold changes that coul…
- arXiv cs.AIresearchREACT: Representational Early Analysis and Control of Chain-of-Thought Trajectories
arXiv:2610.32113v1 Announce Type: cross Abstract: Chain-of-thought (CoT) generation offers an opportunity to intervene before a reasoning model completes its answer. But does the point at which risk is easiest to detect…
- arXiv cs.AIresearchTampered ML: Evaluating LLM Judgment of Machine Learning Research
arXiv:2610.32114v1 Announce Type: cross Abstract: One approach to accelerating progress on AI alignment is to automate parts of alignment research itself. Doing so requires systems that can perform reliably on hard-to-su…
- arXiv cs.AIresearchBeyond Data Schedule: Fast LLM Pretraining on Data Mixtures by Loss-Reweighting
arXiv:2610.32141v1 Announce Type: cross Abstract: Large language models often acquire specialized capabilities through successive pretraining stages, but these gains can come at the cost of forgetting earlier knowledge.…
- arXiv cs.AIresearchRethinking LoRA Initialization for Robust Asymmetric Learning Rates
arXiv:2610.32142v1 Announce Type: cross Abstract: Low-Rank Adaptation (LoRA) is often highly sensitive to learning-rate (LR) choices. Recent empirical studies suggest that, once LRs are carefully tuned, vanilla LoRA can…
- arXiv cs.AIresearchImplicit Neural Representations for Variational Problems on Graphons
arXiv:2610.32147v1 Announce Type: cross Abstract: We propose a neural exploratory framework for variational problems over graphons, i.e., symmetric measurable functions $[0,1]^2 \to [0,1]$, that arise as limits of dense…
- arXiv cs.AIresearchSparsity-driven emergence of functional modularity in multitask neural networks
arXiv:2610.32153v1 Announce Type: cross Abstract: Neural networks can learn multiple tasks, but how learning organizes shared computation into specialized and reusable components remains poorly understood. In this work,…
- arXiv cs.AIresearchMG-REPA: Multi-Granularity Representation Alignment for Generative Target Speaker Extraction
arXiv:2610.32161v1 Announce Type: cross Abstract: Generative target speaker extraction (TSE) models achieve high-quality extraction, but suffer from considerably slow training convergence. Representation Alignment (REPA)…
- arXiv cs.AIresearchSearching with Sparse Histories: Cross-Task Architecture Search from Incomplete Search Logs and Zero-Cost Proxies
arXiv:2610.32179v1 Announce Type: cross Abstract: Neural architecture search (NAS) is expensive, but cross-task transfer can reduce costs by reusing knowledge from previous tasks. However, existing methods struggle with…