AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchEvidence Acquisition and Stopping via Lattice Distillation
arXiv:2606.09433v2 Announce Type: replace Abstract: Sequential evidence acquisition requires deciding what to observe and when to stop from partial information. We introduce a complete-record view of policy supervision:…
- arXiv cs.AIresearchFrame-Conditioned Moral Computation in LLaMA 3.1-8B-Instruct: An Exploratory Mechanistic Interpretability Audit of Ethical Reasoning
arXiv:2606.15507v2 Announce Type: replace Abstract: This exploratory observational study uses the Transluce platform to analyze the internal activity of Llama 3.1-8B-Instruct on 54 moral prompts in four batteries: dilemm…
- arXiv cs.AIresearchScientific discovery as meta-optimization: a combinatorial optimization case study
arXiv:2606.26728v2 Announce Type: replace Abstract: Scientific discovery is fundamentally an optimization problem, defined by a vast ``state space'' of theories and experiments, and an evaluation criterion based on quali…
- arXiv cs.AIresearchArena-T2I Hard: Benchmarking and Improving Faithfulness with Dependency-Aware Checklist
arXiv:2606.31711v2 Announce Type: replace Abstract: Faithfulness -- how precisely a generated image aligns with its prompt -- is increasingly central to the real-world utility of text-to-image (T2I) models. Existing fait…
- arXiv cs.AIresearchHarnessing Textual Refusal Directions for Multimodal Safety
arXiv:2606.31876v3 Announce Type: replace Abstract: Safety in Multimodal Large Language Models (MLLMs) depends on unsafe multimodal data for post-training alignment or activation steering. However, such data is hard to c…
- arXiv cs.AIresearchCan We Trust Item Response Theory for AI Evaluation?
arXiv:2607.15190v3 Announce Type: replace Abstract: AI benchmarks increasingly leverage item-level statistical models, particularly item response theory (IRT), to estimate capabilities, rank systems, select informative e…
- arXiv cs.AIresearchNeurOWL: An LLM-Based Neural-symbolic Framework for Incomplete OWL Ontology Reasoning
arXiv:2607.15776v3 Announce Type: replace Abstract: OWL ontologies provide a formal knowledge representation framework that enables semantic reasoning, and have been widely adopted across domains such as healthcare and b…
- arXiv cs.AIresearchMulti-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
arXiv:2607.22583v3 Announce Type: replace Abstract: Large Language Models (LLMs) have achieved widespread adoption because of their strong reasoning and query-response capabilities. However, deploying them in embedded an…
- arXiv cs.AIresearchCAPT: A Multi-task Continuous Autoregressive Transformer enabling Cross-dataset and Cross-species Transfer for Calcium Population Dynamics
arXiv:2607.23258v3 Announce Type: replace Abstract: Calcium imaging has enabled large-scale recordings of neural population activity across diverse animals, laboratories, and experimental paradigms. However, these datase…
- arXiv cs.AIresearchDefense via Consensus: Mitigating Subliminal Learning through Multi-Teacher Distillation
arXiv:2607.23394v2 Announce Type: replace Abstract: Fine-tuning can introduce unwanted behaviors through deliberate poisoning or unexpected generalization from seemingly benign data. Because these effects may be difficul…
- arXiv cs.AIresearchWIDE: Boosting Adaptive LLM Inference via Token-level Dynamic Width Pruning
arXiv:2607.28418v2 Announce Type: replace Abstract: Pruning is a promising approach for improving the efficiency of large language models (LLMs). Static structured pruning is hardware-friendly, but its input-agnostic com…
- arXiv cs.AIresearchWhen the Judge Should Not Decide: Evidence-Locked, Non-Compensatory Selection Bounds LLM-Judge Failure in Reasoning Pipelines
arXiv:2608.07813v2 Announce Type: replace Abstract: An LLM judge deployed inside a reasoning pipeline does not merely measure quality, it decides which answer ships. We show that the cost of that decision depends less on…
- arXiv cs.AIresearchVERDI:Verified Contradictions for Continual World-Model Optimization
arXiv:2608.09537v2 Announce Type: replace Abstract: Foundation world models have made remarkable progress in planning, simulation, and embodied intelligence. However, optimizing a pretrained world model toward a user-spe…
- arXiv cs.AIresearchThe edge-based contiguous p-median problem: exact formulations and strengthening inequalities
arXiv:2608.11230v2 Announce Type: replace Abstract: This paper introduces the edge-based contiguous p-median (ECpM) problem, which partitions the roads in a network into a given number of compact and contiguous territori…
- arXiv cs.AIresearchBuy the Rumor, Sell the News: When Is News Priced In?
arXiv:2608.14014v2 Announce Type: replace Abstract: Two old market sayings hold that news is already priced in by the time it is published, and that the rumor is bought while the news is sold. Both are empirical claims a…
- arXiv cs.AIresearchWhen Entropy Is Not Enough: Reclaiming Lost Semantics in LLM Output Length Prediction
arXiv:2608.15592v3 Announce Type: replace Abstract: Efficient LLM serving is often bottlenecked by the need to pad sequences to a fixed maximum length, and this wastes compute and degrades throughput. Output length predi…
- arXiv cs.AIresearch6.5% of the Neuro-Symbolic Literature Can Be Reproduced from Its Published Artifacts, a Six-Stage Audit Framework and First Instantiation
arXiv:2608.26236v2 Announce Type: replace Abstract: We present a six-stage framework for auditing the reproducibility of scientific claims across a research literature within the computer science domain, and instantiate…
- arXiv cs.AIresearchOmni-Interactive Universal Embedder
arXiv:2608.27044v2 Announce Type: replace Abstract: Multimodal representation learning has been shifting from traditional two-tower architectures to large language model (LLM)-based embedders due to their strong instruct…
- arXiv cs.AIresearchMultimodal Resource-Exhaustion Attacks on Vision-Language Models via Joint Pixel-Prompt Optimization
arXiv:2609.05889v2 Announce Type: replace Abstract: Resource-exhaustion attacks against autoregressive vision-language models (VLMs) typically assume unimodal threat models, treating the image branch as the primary optim…
- arXiv cs.AIresearchApplication of curiosity driven exploration methods for hardware interference identification
arXiv:2609.08729v2 Announce Type: replace Abstract: The transition from single-core to multi-core architectures in safety-critical embedded systems introduces significant challenges due to inter-core interference caused…
- arXiv cs.AIresearchDifferential Functionalism: Jacobians Shape the Structure of Phenomenal Experience
arXiv:2609.09306v3 Announce Type: replace Abstract: This paper investigates the hypothesis that the first-order structure of physical interactions (Jacobians) shapes the structure of phenomenal experience. Two measures o…
- arXiv cs.AIresearchHow Good Are Frontier Models at Physics? Expert Re-Grading Reveals Broken Evaluations and Near-Saturation of Leading Benchmarks
arXiv:2609.13009v2 Announce Type: replace Abstract: Low reported scores on leading physics benchmarks, including those featured in the Artificial Analysis Intelligence Index (2026), suggest that frontier language models…
- arXiv cs.AIresearchGradient Boosting Reinforcement Learning
arXiv:2407.08250v3 Announce Type: replace-cross Abstract: We present Gradient Boosting Reinforcement Learning (GBRL), a framework that adapts the strengths of gradient boosting trees (GBT) to reinforcement learning (RL)…
- arXiv cs.AIresearchMitigating Data Reconstruction Attacks in Federated Learning: Conditions, Bounds, and Theoretical Guarantees
arXiv:2407.16735v2 Announce Type: replace-cross Abstract: Privacy attacks, particularly data reconstruction attacks, pose significant challenges to federated learning despite its promise in privacy-preserving collaborati…
- arXiv cs.AIresearchForecast-PEFT: Parameter-Efficient Fine-Tuning for Pre-trained Motion Forecasting Models
arXiv:2407.19564v2 Announce Type: replace-cross Abstract: Recent advancements in motion forecasting have been significantly accelerated by self-supervised pre-training techniques. However, current methods typically rely…
- arXiv cs.AIresearchCompetition and Diversity in Generative AI
arXiv:2412.08610v4 Announce Type: replace-cross Abstract: Recent evidence, both in the lab and in the wild, suggests that the use of generative artificial intelligence reduces the diversity of content produced. The use o…
- arXiv cs.AIresearchReal Faults in Deep Learning Fault Benchmarks: How Real Are They?
arXiv:2412.16336v2 Announce Type: replace-cross Abstract: As the adoption of Deep Learning (DL) systems continues to rise, an increasing number of approaches are being proposed to test these systems, localise faults with…
- arXiv cs.AIresearchPDB: Not All Drivers Are the Same -- A Personalized Dataset for Understanding Driving Behavior
arXiv:2503.06477v2 Announce Type: replace-cross Abstract: Driving behavior varies across individuals, yet most existing driving datasets do not preserve driver identity or control for factors that may affect driving styl…
- arXiv cs.AIresearchTestDG: Test-time Domain Generalization for Continual Test-time Adaptation
arXiv:2504.04981v4 Announce Type: replace-cross Abstract: This paper studies continual test-time adaptation (CTTA), the task of adapting a model to constantly changing unseen domains in testing while preserving previousl…
- arXiv cs.AIresearchStudent Perspectives on the Benefits and Risks of AI in Education
arXiv:2505.02198v3 Announce Type: replace-cross Abstract: The use of chatbots equipped with artificial intelligence (AI) in educational settings has increased in recent years, showing potential to support teaching and le…
- arXiv cs.AIresearchTo Trust Or Not To Trust Your Vision-Language Model's Prediction
arXiv:2505.23745v3 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) have demonstrated strong capabilities in aligning visual and textual modalities, enabling a wide range of applications in multimodal…
- arXiv cs.AIresearchUNO: Unlearning via Orthogonalization in Generative models
arXiv:2506.04712v3 Announce Type: replace-cross Abstract: As generative models become increasingly powerful and pervasive, the ability to unlearn specific data due to privacy concerns, legal requirements, or harmful cont…
- arXiv cs.AIresearchChain of Retrieval: Multi-Aspect Iterative Search Expansion and Post-Order Search Aggregation for Full Paper Retrieval
arXiv:2507.10057v4 Announce Type: replace-cross Abstract: Scientific paper retrieval, particularly framed as document-to-document retrieval, aims to identify relevant papers in response to a long-form query paper, rather…
- arXiv cs.AIresearchNLP Meets the World: Toward Improving Conversations With News Media About Natural Language Processing Research
arXiv:2507.10559v3 Announce Type: replace-cross Abstract: Recent advances in large language models (LLMs) have rapidly increased the public's interest in natural language processing (NLP). This attention is reflected by…
- arXiv cs.AIresearchDynamic gain neuromodulation attenuates stability gaps during continual joint training
arXiv:2507.14056v4 Announce Type: replace-cross Abstract: Recent work in continual learning has highlighted the stability gap -- a temporary performance drop on previously learned tasks when new ones are introduced. This…
- arXiv cs.AIresearchEquivariant Volumetric Grasping
arXiv:2507.18847v4 Announce Type: replace-cross Abstract: We propose a new volumetric grasp model that is equivariant to rotations around the vertical axis, leading to a significant improvement in sampling efficiency. Ou…
- arXiv cs.AIresearchSlotMatch: Distilling Object-Centric Representations for Unsupervised Video Segmentation
arXiv:2508.03411v4 Announce Type: replace-cross Abstract: Unsupervised video segmentation is a challenging computer vision task, especially due to the lack of supervisory signals coupled with the complexity of visual sce…
- arXiv cs.AIresearchOpen-sci-ref-0.02: open reference baselines for reproducible language model and dataset comparison
arXiv:2509.09009v4 Announce Type: replace-cross Abstract: Reliable progress in language modeling depends on knowing whether a new dataset, architecture, or training recipe is genuinely better than existing alternatives.…
- arXiv cs.AIresearchToward a Physics of Deep Learning and Brains
arXiv:2509.22649v2 Announce Type: replace-cross Abstract: Deep neural networks and brains both learn and share superficial similarities: processing nodes are likened to neurons and adjustable weights are likened to modif…
- arXiv cs.AIresearchVeriGRAG: Enhancing LLM-Based Verilog Code Generation with Structure-Aware Soft Prompts
arXiv:2510.15914v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have demonstrated remarkable capabilities in generating Verilog code from natural language descriptions. However, Verilog code inhere…
- arXiv cs.AIresearchThe Ends Justify the Thoughts: RL-Induced Motivated Reasoning Fools CoT Monitors
arXiv:2510.17057v3 Announce Type: replace-cross Abstract: Chain-of-thought (CoT) monitoring has emerged as a compelling method for detecting harmful behaviors such as reward hacking in reasoning models, under the assumpt…
- arXiv cs.AIresearchManifold Approximation leads to Robust Kernel Alignment
arXiv:2510.22953v2 Announce Type: replace-cross Abstract: Centered kernel alignment (CKA) is a popular metric for comparing representations, determining equivalence of networks, and conducting neuroscience research. Howe…
- arXiv cs.AIresearchEngTrace: A Symbolic Benchmark for Verifiable Process Evaluation of Engineering Reasoning
arXiv:2511.01650v4 Announce Type: replace-cross Abstract: Most engineering problems involve complex multi-step derivations through checkable intermediate quantities. Grading only the final answer misjudges LLMs, creditin…
- arXiv cs.AIresearchA Privacy-Preserving Federated Learning Method for Heterogeneous Clients
arXiv:2511.06064v2 Announce Type: replace-cross Abstract: Federated Learning (FL) with Differential Privacy (DP) can ensure the protection of healthcare data privacy against malicious user attacks. However, FL with the D…
- arXiv cs.AIresearchContinually Evolving Skill Knowledge in Vision Language Action Model
arXiv:2511.18085v5 Announce Type: replace-cross Abstract: Vision-language-action (VLA) models show promising knowledge accumulation ability through pretraining, yet continual learning remains challenging. Existing contin…
- arXiv cs.AIresearchCerium: A Multi-GPU Framework for Terabyte-Scale Encrypted Inference
arXiv:2512.11269v2 Announce Type: replace-cross Abstract: Encrypted AI using fully homomorphic encryption (FHE) enables inference directly over encrypted queries, providing strong privacy guarantees. However, its computa…
- arXiv cs.AIresearchIRPM: Intergroup Relative Preference Modeling for Pointwise Generative Reward Models
arXiv:2601.00677v3 Announce Type: replace-cross Abstract: Generative Reward Models (GRMs) have demonstrated strong performance in reward modeling, due to their interpretability and potential for refinement through reinfo…
- arXiv cs.AIresearchSteering Language Models Before They Speak: Logit-Level Interventions
arXiv:2601.10960v3 Announce Type: replace-cross Abstract: Controllable generation requires language models to realize output characteristics such as reading level, politeness, and toxicity. Existing steering methods are…
- arXiv cs.AIresearchPatch-Level Tokenization with CNN Encoders and Attention for Improved Transformer Time-Series Forecasting
arXiv:2601.12467v4 Announce Type: replace-cross Abstract: Transformer-based models have shown strong performance in time-series forecasting by leveraging self-attention to model long-range temporal dependencies. However,…
- arXiv cs.AIresearchTest-Time Compute Games
arXiv:2601.21839v3 Announce Type: replace-cross Abstract: Test-time compute has emerged as a promising strategy to enhance the reasoning abilities of large language models (LLMs). However, this strategy has in turn incre…