AI news archive.
10,660 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchFlexQuant: Elastic Quantization Framework for Locally Hosted LLM on Edge Devices
arXiv:2501.07139v2 Announce Type: replace Abstract: Deploying LLMs on edge devices presents serious technical challenges. Memory elasticity is crucial for edge devices with unified memory, where memory is shared and fluc…
- arXiv cs.AIresearchFrom Reasoning to Generalization: Knowledge-Augmented LLMs for ARC Benchmark
arXiv:2505.17482v2 Announce Type: replace Abstract: Recent reasoning-oriented LLMs have demonstrated strong performance on challenging tasks such as mathematics and science examinations. However, core cognitive faculties…
- arXiv cs.AIresearchSearching for Actual Causes: Approximate Algorithms with Adjustable Precision
arXiv:2507.07857v2 Announce Type: replace Abstract: Causality has gained increasing attention in recent years, notably for improving the interpretability of machine learning models. Yet the field of explainable artificia…
- arXiv cs.AIresearchRethinking Prospect Theory for LLMs: Revealing the Instability of Decision-Making under Epistemic Uncertainty
arXiv:2508.08992v5 Announce Type: replace Abstract: Real-world decision-making often involves uncertainty expressed in linguistic rather than numerical terms, and Prospect Theory (PT) provides a classic framework for mod…
- arXiv cs.AIresearchAI-driven ionic liquid discovery with unified chemical intelligence
arXiv:2511.11257v2 Announce Type: replace Abstract: The rational discovery of functional solvents remains challenging because macroscopic solvent behaviors emerge from complex many-body interactions across vast chemical…
- arXiv cs.AIresearchWhat If TSF: Reframing Time Series Forecasting as Scenario-Guided Multimodal Forecasting
arXiv:2601.08509v2 Announce Type: replace Abstract: Recent advances in large language models (LLMs) have enabled time series forecasting to move beyond numerical observations and incorporate external information in multi…
- arXiv cs.AIresearchSTEP-LLM: Generating CAD STEP Models from Natural Language with Large Language Models
arXiv:2601.12641v2 Announce Type: replace Abstract: Computer-aided design (CAD) is vital to modern manufacturing, yet model creation remains labor-intensive and expertise-heavy. To enable non-experts to translate intuiti…
- arXiv cs.AIresearchGLOVE: Global Verifier for LLM Memory-Environment Realignment
arXiv:2601.19249v2 Announce Type: replace Abstract: Most existing memory-enhanced Large Language Model (LLM) approaches implicitly assume that memory validity can be established either through external evaluators that pr…
- arXiv cs.AIresearchBayesian-LoRA: Probabilistic Low-Rank Adaptation of Large Language Models
arXiv:2601.21003v4 Announce Type: replace Abstract: Large Language Models usually put more emphasis on accuracy and therefore, will guess even when not certain about the prediction, which is especially severe when fine-t…
- arXiv cs.AIresearchLatent Chain-of-Thought as Planning: Decoupling Reasoning from Verbalization
arXiv:2601.21358v3 Announce Type: replace Abstract: Chain-of-Thought (CoT) empowers Large Language Models (LLMs) to tackle complex problems, but remains constrained by the computational cost and early token commitments i…
- arXiv cs.AIresearchDo Latent-CoT Models Think Step-by-Step? A Mechanistic Study on Sequential Reasoning Tasks
arXiv:2602.00449v2 Announce Type: replace Abstract: Latent Chain-of-Thought aims to enable step-by-step computation without emitting long rationales, yet its internal mechanisms remain unclear. We study CODI, a continuou…
- arXiv cs.AIresearchAdversarial Reward Auditing for Active Detection and Mitigation of Reward Hacking
arXiv:2602.01750v2 Announce Type: replace Abstract: Reinforcement Learning from Human Feedback (RLHF) remains vulnerable to reward hacking, where models exploit spurious correlations in learned reward models to achieve h…
- arXiv cs.AIresearchKANFIS: A Neuro-Symbolic Framework for Interpretable and Uncertainty-Aware Learning
arXiv:2602.03034v3 Announce Type: replace Abstract: Adaptive Neuro-Fuzzy Inference System (ANFIS) was designed to combine the learning capabilities of neural network with the reasoning transparency of fuzzy logic. Howeve…
- arXiv cs.AIresearchSemantic Purification for Conditional Representation Learning
arXiv:2602.05464v2 Announce Type: replace Abstract: Conditional representation learning aims to extract criterion-specific features for customized tasks. Recent methods construct conditional subspaces spanned by criterio…
- arXiv cs.AIresearchSelective Fine-Tuning for Targeted and Robust Concept Unlearning
arXiv:2602.07919v2 Announce Type: replace Abstract: Text guided diffusion models are used by millions of users, but can be easily exploited to produce harmful content. Concept unlearning methods aim at reducing the model…
- arXiv cs.AIresearchGUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments
arXiv:2602.09007v3 Announce Type: replace Abstract: An image generation model used as a graphical user interface (GUI) environment must produce not only a plausible screen but also the correct response to an action. Exis…
- arXiv cs.AIresearchCausalReasoningBenchmark: A Real-World Benchmark for Disentangled Evaluation of Causal Identification and Estimation
arXiv:2602.20571v3 Announce Type: replace Abstract: Many benchmarks for automated causal inference evaluate a system's performance based on a single numerical output, such as an Average Treatment Effect (ATE). This appro…
- arXiv cs.AIresearchWords & Weights: Streamlining Multi-Turn Interactions via Co-Adaptation
arXiv:2603.01375v2 Announce Type: replace Abstract: Test-time policy adaptation for multi-turn interactions (T2PAM) is essential for aligning Large Language Models (LLMs) with dynamic user needs during inference time. Ho…
- arXiv cs.AIresearchEveryQuery: A Promptable Foundation Model for Clinical Prediction Tasks over Electronic Health Records
arXiv:2603.07900v3 Announce Type: replace Abstract: Autoregressive foundation models for electronic health records can make predictions for diverse clinical tasks without finetuning. These methods work by generating poss…
- arXiv cs.AIresearchLearning Transferable Sensor Models via Language-Informed Pretraining
arXiv:2603.11950v2 Announce Type: replace Abstract: Multimodal language models have demonstrated strong semantic understanding and reasoning over physiological and behavioral signals captured from diverse healthcare sens…
- arXiv cs.AIresearchLearning Dynamic Belief Graphs for Theory-of-mind Reasoning
arXiv:2603.20170v2 Announce Type: replace Abstract: Theory of Mind (ToM) reasoning with Large Language Models (LLMs) requires inferring how people's implicit, evolving beliefs shape what they seek and how they act under…
- arXiv cs.AIresearchMonitorBench: A Comprehensive Benchmark for Chain-of-Thought Monitorability in Large Language Models
arXiv:2603.28590v4 Announce Type: replace Abstract: Large language models (LLMs) can generate chains of thought (CoTs) that are not always causally responsible for their final outputs. When such a mismatch occurs, the Co…
- arXiv cs.AIresearchMeasuring (some aspects of) the metacognition of AI
arXiv:2603.29693v4 Announce Type: replace Abstract: A robust decision-making process must take into account uncertainty, especially when the choice involves inherent risks. Because artificial intelligence (AI) systems ar…
- arXiv cs.AIresearchRefineRL: Advancing Competitive Programming with Self-Refinement Reinforcement Learning
arXiv:2604.00790v2 Announce Type: replace Abstract: While large language models (LLMs) have demonstrated strong performance on complex reasoning tasks such as competitive programming (CP), existing methods predominantly…
- arXiv cs.AIresearchAcceptance Dynamics Across Cognitive Domains in Speculative Decoding
arXiv:2604.14682v2 Announce Type: replace Abstract: Speculative decoding accelerates large language model (LLM) inference. It uses a small draft model to propose a tree of future tokens. A larger target model then verifi…
- arXiv cs.AIresearchHindsight Compacts but Does Not Repair: Rethinking On-Policy Self-Distillation in Reasoning Models
arXiv:2605.06188v2 Announce Type: replace Abstract: On-Policy Self-Distillation (OPSD) has emerged as a promising post-training method: requiring only privileged context, it enables token-level credit assignment from a s…
- arXiv cs.AIresearchHow Deep Can LLMs Learn to Reason? Expressiveness Is Key
arXiv:2605.06638v4 Announce Type: replace Abstract: Reinforcement learning (RL) has been applied to improve large language model (LLM) reasoning, yet the systematic study of how training scales with task difficulty has b…
- arXiv cs.AIresearchSearchSkill: Teaching LLMs to Use Search Tools with Evolving Skill Banks
arXiv:2605.09038v4 Announce Type: replace Abstract: Teaching language models to use search tools is not only a question of whether they search, but also of whether they issue good queries. This is especially important in…
- arXiv cs.AIresearchHow LLMs Are Persuaded: A Few Attention Heads, Rerouted
arXiv:2605.09314v2 Announce Type: replace Abstract: Large language models can answer a factual question correctly, yet switch to an incorrect answer when exposed to persuasive context. Why this happens inside the model r…
- arXiv cs.AIresearchPACE: Policy-Native Adaptive Decision Timing for Long-Horizon Reasoning
arXiv:2605.09860v5 Announce Type: replace Abstract: Long-horizon reasoning requires deciding not only what actions to take, but how many to execute open-loop before replanning. This number, the execution depth, balances…
- arXiv cs.AIresearchDistribution-Aware Programming: Learning Specialized Solvers from Experience
arXiv:2605.14141v3 Announce Type: replace Abstract: Many optimization problems are solved repeatedly on instances drawn from the same underlying distribution. In this setting, a system can begin with a general-purpose so…
- arXiv cs.AIresearchUniversal Quantum Transformer
arXiv:2606.00045v4 Announce Type: replace Abstract: Classical continuous-space neural networks fundamentally struggle to lock into exact formal rules, whether mathematical, such as modular arithmetic and non-Abelian grou…
- arXiv cs.AIresearchCapability Self-Assessment in Large Language Models
arXiv:2606.00251v2 Announce Type: replace Abstract: The ability to recognize one's own limitations and decide whether to solve a problem or seek help is fundamental to reliable intelligent systems. Yet modern large langu…
- arXiv cs.AIresearchNovelAPIBench: Diagnosing How A Code LLM Learns to Use Novel APIs
arXiv:2606.03657v2 Announce Type: replace Abstract: Rapidly evolving software libraries require coding language models to use unfamiliar APIs. Yet failed solutions alone tell us little about what API knowledge models nee…
- arXiv cs.AIresearchMultilingual Fine-Tuning via Localized Gradient Conflict Resolution
arXiv:2606.05613v2 Announce Type: replace Abstract: Cross-lingual versatility has become a defining capability of Large Language Models (LLMs). However, fine-tuning these models often induces negative interference, where…
- arXiv cs.AIresearchSafety Paradox: How Enhanced Safety Awareness Leaves LLMs Vulnerable to Posterior Attack
arXiv:2606.05614v2 Announce Type: replace Abstract: Large language models (LLMs) are rigorously aligned to refuse harmful requests, a process that inherently cultivates a latent capacity to evaluate and recognize unsafe…
- arXiv cs.AIresearchMaking AI Scientists Auditable from Evidence to Claim
arXiv:2606.18874v4 Announce Type: replace Abstract: AI research systems can report improved results without establishing that the tested implementation matches the method described in their conclusions. We introduce Xcie…
- arXiv cs.AIresearchCliff Tokens: Analyzing Failure Trigger Tokens in LLM Mathematical Reasoning
arXiv:2606.25524v3 Announce Type: replace Abstract: Large language models reach high accuracy in mathematical reasoning, but individual traces on the same problem diverge; some arrive at the correct answer while others f…
- arXiv cs.AIresearchTandem Reinforcement Learning with Verifiable Rewards
arXiv:2606.28166v2 Announce Type: replace Abstract: Reinforcement learning with verifiable rewards (RLVR) has driven substantial gains in language-model reasoning. However, models fine-tuned with RLVR need not remain com…
- arXiv cs.AIresearchMechanistic Personality Analysis of LLMs: Steering Personality via Latent Feature Interventions
arXiv:2606.28770v2 Announce Type: replace Abstract: Large Language Models (LLMs) have demonstrated the ability to simulate human-like OCEAN personality traits in generated text. Previous efforts have focused on prompt en…
- arXiv cs.AIresearchFacePlex: Toward Natural Full-Duplex Conversational Avatars
arXiv:2606.30145v2 Announce Type: replace Abstract: Natural human conversation is inherently a real-time interaction in which speech and facial behavior continuously evolve. Enabling such interaction requires a conversat…
- arXiv cs.AIresearchWhat is Left for Us? Second Scholarship Against the Degradation of Research by AI
arXiv:2607.04049v3 Announce Type: replace Abstract: We argue that generative AI can degrade research by eroding the very practices through which scholarly judgement is formed. As constitutive conditions for the productio…
- arXiv cs.AIresearchTowards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in Large Language Model Finetuning
arXiv:2607.08393v2 Announce Type: replace Abstract: Fine-tuning LLMs to inject new knowledge faces a critical challenge: LLMs can quickly memorize new facts, yet fail to use them for downstream reasoning tasks. We formal…
- arXiv cs.AIresearchThe Steering Budget: Examples beat Knobs
arXiv:2607.14246v2 Announce Type: replace Abstract: Generative models are steered with knobs -- prompts, guidance scales, property tags. Turn one as hard as you like and, past a point, it stops moving the property you ca…
- arXiv cs.AIresearchSemantically Similar, Yet Not Answerable: Diagnosing the Semantic-Answerability Gap in Table RAG
arXiv:2607.17742v2 Announce Type: replace Abstract: In retrieval-augmented generation (RAG), semantic relevance asks whether a source matches a query in meaning, while answerability asks whether it contains sufficient in…
- arXiv cs.AIresearchSAGE: Subgoal-Conditioned Action Generation for Latent World Model Planning
arXiv:2607.17973v2 Announce Type: replace Abstract: Latent world models have emerged as a powerful planning paradigm by learning action-conditioned predictive dynamics and using them as internal simulators to imagine and…
- arXiv cs.AIresearchDWM: Separating World Effects from Actions in Latent World Models
arXiv:2607.18715v2 Announce Type: replace Abstract: Latent world models underpin much of modern model-based control, yet current action-conditioned formulations supervise the next-latent transition with a single, undiffe…
- arXiv cs.AIresearchCodifying the Judge: Scalable Evaluation via Program Distillation
arXiv:2607.22561v2 Announce Type: replace Abstract: LLM-as-a-judge has become the standard for automated evaluation, but it suffers from high cost, inference latency, and opaque decisions---limitations that undermine its…
- arXiv cs.AIresearchHow LLM Task-Adaptation Reshapes Alignment: A Multi-dimensional Study of Behavioral and Representational Drift
arXiv:2607.22676v2 Announce Type: replace Abstract: Post-training is a key mechanism for adapting large language models to downstream tasks. While prior work suggests that task adaptation can alter a model's pre-existing…
- arXiv cs.AIresearchMissClick: Execution-Aware Adversarial Attacks on Coordinate Generation in GUI Grounding Models
arXiv:2608.03740v2 Announce Type: replace Abstract: Recent GUI visual grounding models generate screen coordinates as digit-token sequences that are parsed into numerical values and mapped to executable clicks. This gene…