AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchConcept-Grounded Attention: A Controlled Evaluation of Graph-Injected Attention, Temporal Versioning, and Epistemic Status
arXiv:2609.38684v1 Announce Type: new Abstract: Knowledge-intensive language-model systems typically represent external knowledge as text chunks or static graphs, with limited support for concept evolution, point-in-time…
- arXiv cs.AIresearchGATE-ST: Gene-Aware Text-image Encoder for Spatial Transcriptomics
arXiv:2609.38690v1 Announce Type: new Abstract: Spatial transcriptomics enables spatially resolved gene expression analysis from slide-level images while preserving morphological features, providing valuable information…
- arXiv cs.AIresearchBudget Boundary Effects in Test-Time Mathematical Reasoning
arXiv:2609.38699v1 Announce Type: new Abstract: A cumulative token cap can fall inside a mathematical derivation, forcing a test-time controller to choose between stopping at the cap (strict) and allowing the current att…
- arXiv cs.AIresearchUniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement
arXiv:2609.38721v1 Announce Type: new Abstract: Modern multimodal models bring generation and understanding into a single unified system, which enables them to provide and learn from their own feedback. Motivated by this…
- arXiv cs.AIresearchCode to Control: Synthesizing Parameterized Reactive Controllers
arXiv:2609.38733v1 Announce Type: new Abstract: Recent LLM-based approaches to control either invoke a language model to select actions or synthesize world models that require planning at every decision, introducing late…
- arXiv cs.AIresearchWhen Reasoning Goes Astray: Attention Dynamics of Uncontrolled Reasoning
arXiv:2609.38817v1 Announce Type: new Abstract: Large reasoning models (LRMs) improve performance on complex tasks through extended reasoning, yet the same process can degenerate into redundant verification and persisten…
- arXiv cs.AIresearchWhose Voice Survives the Summary? A Voice-Retention Audit of LLM Employee Listening
arXiv:2609.38818v1 Announce Type: new Abstract: Organizations increasingly route employee feedback to leaders through large language model (LLM) summaries, an unaudited layer that silences already-spoken voice. We introd…
- arXiv cs.AIresearchMore Choices, Fewer Decisions: Ordinal-Scale Bias in JEV-like Direct-Decision Models
arXiv:2609.38827v2 Announce Type: replace Abstract: Direct-decision models enable efficient classification and evaluation by selecting from predefined candidates, but their reliability depends on more than prediction acc…
- arXiv cs.AIresearchDiversity Combining for Multi-Path LLM Reasoning
arXiv:2609.38829v2 Announce Type: replace Abstract: Multi-path reasoning methods such as self-consistency (SC) sample $K$ reasoning paths and choose the most frequent answer. However, their gains quickly plateau as $K$ i…
- arXiv cs.AIresearchOpenJev-RLCD: A Working RLCD Implementation
arXiv:2609.38850v1 Announce Type: new Abstract: Decision models such as Jev answer questions with probabilities, which are only useful if they are calibrated. Open-source reproductions rely on supervised fine-tuning plus…
- arXiv cs.AIresearchReasoning Externalization for Faithful Large Language Model Narratives of Stock Return Predictions
arXiv:2609.38869v2 Announce Type: replace Abstract: In finance, interpreting machine learning predictions is essential, yet the numerical outputs of explainable AI can be difficult for non-experts to understand. While la…
- arXiv cs.AIresearchHow Much Can Reliability Drift Under a Fixed Confidence Distribution?
arXiv:2609.38917v1 Announce Type: new Abstract: A classifier's conditional accuracy can change while its confidence distribution stays exactly the same. We study the worst-case movement of the reliability relation under…
- arXiv cs.AIresearchPrototype-guided Bilateral Alignment Multimodal Federated Learning
arXiv:2609.38925v1 Announce Type: new Abstract: Multimodal federated learning (MFL) has emerged as a pivotal paradigm for leveraging distributed data to enhance model performance. However, existing methods predominantly…
- arXiv cs.AIresearchLearning What to Forget: Distributional Unlearning for LLM Representation Spaces
arXiv:2609.38929v1 Announce Type: new Abstract: Machine learning systems increasingly face the need to remove the influence of entire data domains, such as toxic language, harmful behavior, or topical content, rather tha…
- arXiv cs.AIresearchRouting Probes Can Improve Without New Information: An Exact-Null Audit of Uncertainty Beyond Model Outputs
arXiv:2609.38956v1 Announce Type: new Abstract: Routing signals of modern vision transformers -- expert gates, attention-residual weights and halting scores -- often improve probes that predict whether the model is corre…
- arXiv cs.AIresearchTargeted Retrieval, Compact Representations: How CoT Reasoning Improves Long-Context Counting
arXiv:2609.38958v1 Announce Type: new Abstract: Large language models (LLMs) have been rapidly improving in long-context tasks, powered by Chain-of-Thought (CoT) reasoning. However, the internal mechanisms underlying thi…
- arXiv cs.AIresearchAlleviating Hallucination in Reasoning Tasks with Training-Free Uncertainty-Guided Steering
arXiv:2609.38962v1 Announce Type: new Abstract: Recent work on hallucination detection in large language models has shown that, for a fixed pre-trained model and reasoning task, it is possible to estimate the model's con…
- arXiv cs.AIresearchC-STRIDE: An Observation-Driven AI Digital Twin for Predicting Basin-Wide Flood Fields from Sparse Stream-Gauge Histories
arXiv:2609.39005v1 Announce Type: new Abstract: Emergency managers need to know where floodwater is, how deep it is, and how it will change over the coming hours across an entire river basin. During a flood, however, rea…
- arXiv cs.AIresearchMulti-LLM Collaborative Alignment via Stackelberg Games
arXiv:2609.39076v2 Announce Type: replace Abstract: A pool of language models can collaborate and improve collectively by learning from one another's responses. These interactions depend on the instructions used during t…
- arXiv cs.AIresearchReserve-Aware Contrast Certificates for Conservative Bandits with Uncertain Baselines
arXiv:2609.39106v1 Announce Type: new Abstract: Conservative bandits must improve an incumbent policy without exhausting a prescribed performance budget. When the incumbent is uncertain, separately bounding candidate and…
- arXiv cs.AIresearchBELIEFRAG: Making Adaptive RAG State-Aware under Evolving Evidence
arXiv:2609.39139v1 Announce Type: new Abstract: Adaptive RAG uses signals such as confidence, relevance, support, and retrieval quality to decide when to search or correct evidence. In multi-step retrieval, however, thes…
- arXiv cs.AIresearchReinforcing Multimodal Reasoning via Token-Level Perception-Grounded Advantage Estimation
arXiv:2609.39168v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has improved the reasoning capabilities of Multimodal Large Language Models (MLLMs), yet existing frameworks rely on c…
- arXiv cs.AIresearchLearning Process Rewards via Reasoning State Propagation
arXiv:2609.39220v1 Announce Type: new Abstract: Process reward models (PRMs) have demonstrated notable effectiveness in test-time scaling and reinforcement learning by providing fine-grained signals for evaluating interm…
- arXiv cs.AIresearchEffective Does Not Mean Useful: Conditional Functional Substitutability for Redundancy and Scaling in Transformers
arXiv:2609.39259v1 Announce Type: new Abstract: Modern neural networks scale predictably, yet the mechanisms behind these regularities remain unclear. Neural redundancy is typically characterized by component importance…
- arXiv cs.AIresearchANI: Adaptive Numerical Injection for Unifying Semantic and Arithmetic Representations in Numerical Reasoning
arXiv:2609.39294v1 Announce Type: new Abstract: Precise numerical reasoning with Large Language Models (LLMs) is essential for expanding their applicability to complex real-world tasks. However, text-based tokenization o…
- arXiv cs.AIresearchUnderstanding as No-Arbitrage: Bounded Dutch Books as a Definition and Training Objective for Language Models
arXiv:2609.39341v1 Announce Type: new Abstract: Does a language model merely predict tokens, or does it understand what it says? We make this question measurable by defining "understanding" through the lens of no-arbitra…
- arXiv cs.AIresearchThe Golden Path Hypothesis: Reusable Schedules in Diffusion Caching
arXiv:2609.39343v1 Announce Type: new Abstract: Diffusion caching accelerates generation by replacing transformer computation with cached or predicted features at selected denoising steps. We introduce the Golden Path Hy…
- arXiv cs.AIresearchOn the Complexity of Preference-Based Bandits
arXiv:2609.39351v1 Announce Type: new Abstract: We study preference-based bandits with general reward function classes, where a learner sequentially selects pairs of arms and observes binary preference feedback governed…
- arXiv cs.AIresearchAutoresearch in Mixed-Integer Linear and Nonlinear Programming
arXiv:2609.39360v2 Announce Type: replace Abstract: Despite recent progress in autoresearch, applying it to practical operations research problems, typically formulated as NP-hard mixed-integer linear or nonlinear progra…
- arXiv cs.AIresearchCan Computation from Earlier Problems Help LLMs Solve New Ones?
arXiv:2609.39394v1 Announce Type: new Abstract: Large language models often solve independent problems in the same conversation. Can computation from earlier problems help them solve new ones? To answer this question, we…
- arXiv cs.AIresearchAdvancing Entropy-Level Credit Assignment in RLVR via Proximal Entropy Policy Optimization
arXiv:2609.39402v2 Announce Type: replace Abstract: Value-model-free RLVR methods such as GRPO assign uniform advantages to all tokens in a rollout, ignoring that tokens contribute unequally. Recent methods use token ent…
- arXiv cs.AIresearchInferring Causal Relations between Two Sequences of Events with Language Models
arXiv:2609.39406v1 Announce Type: new Abstract: Causal AI is a branch of Artificial Intelligence which helps understand and reason about cause and effect relationships, not just patterns or correlations. Causal discovery…
- arXiv cs.AIresearchWho Owns That? Evaluating Ownership Intuitions in Large Language Models
arXiv:2609.39483v1 Announce Type: new Abstract: Ownership establishes rights over the use, control, and transfer of objects. Understanding these relations is essential for AI systems to interact appropriately with people…
- arXiv cs.AIresearchDisentangling Self-Distillation: Measuring and Modeling Acquisition and Retention
arXiv:2609.39494v1 Announce Type: new Abstract: Self-distillation with privileged context adapts a language model from demonstrations by letting the model, once conditioned on a reference response, teach its context-free…
- arXiv cs.AIresearchReferential Uncertainty in Human--AI Collaboration
arXiv:2609.39518v1 Announce Type: new Abstract: Effective human-AI collaboration requires partners to establish references through interaction, which becomes fragile when descriptions are ambiguous, similar referents com…
- arXiv cs.AIresearchWhy Do Conventional World Models Fail to Learn Cellular Automata?
arXiv:2609.39604v1 Announce Type: new Abstract: Although conventional world models - auto-regressive or diffusion models based on transformers or convolutional networks - may learn surface statistics of world dynamics, c…
- arXiv cs.AIresearchValues as Style: Disentangling Values from Semantics with One-Way Mixing for Low-Damage LLM Steering
arXiv:2609.39701v1 Announce Type: new Abstract: Value steering should change an LLM's normative priorities while preserving the scenario, facts, and task constraints underlying its answer. Conventional activation edits o…
- arXiv cs.AIresearchA helps B while B hurts A: directed transfer in instruction-tuning mixture
arXiv:2609.39702v1 Announce Type: new Abstract: Adapting a language model to a specialized corpus means choosing which instruction-tuning tasks to train on under a fixed budget, and testing one choice costs a fine-tuning…
- arXiv cs.AIresearchArchitectureIQ: On the Measure of Training Intuition
arXiv:2609.39714v1 Announce Type: new Abstract: Top researchers have good intuition, but do language models have as good intuition about model training as top AI researchers? To measure model intuition of LLMs and humans…
- arXiv cs.AIresearchLearning Steganography Is Easy, Learning Steganographic Reasoning Is Hard
arXiv:2609.39838v1 Announce Type: new Abstract: Chain-of-thought monitoring as an approach for AI oversight and control is threatened by the possibility of steganographic reasoning, where LLMs conceal their reasoning ins…
- arXiv cs.AIresearchFIGS: Evaluating Multi-Turn Sycophancy Without Penalizing Empathy
arXiv:2609.39863v1 Announce Type: new Abstract: Large language models frequently fail to balance staying truthful with being supportive. They often exhibit sycophancy in responses to users, agreeing with false claims, of…
- arXiv cs.AIresearchCompletion-Aware Cross-Fidelity Offline-to-Online Reinforcement Learning for Multi-Line Bus Holding
arXiv:2609.39868v2 Announce Type: replace Abstract: Exploratory reinforcement learning (RL) on an operating bus fleet is impractical,while policies trained only from historical data cannot acquire new experience. Hybrid…
- arXiv cs.AIresearchGrammarRL: Effective Grammar-Constrained Decoding via Reinforcement Learning
arXiv:2609.39869v1 Announce Type: new Abstract: Grammar-constrained generation guarantees syntactic validity, but can substantially degrade semantic quality when the model's preferred outputs are poorly aligned with the…
- arXiv cs.AIresearchConflictGuide: AutoResearch Improves When Competing Behaviors Are Made Visible
arXiv:2609.39933v1 Announce Type: new Abstract: When designing machine learning models, desirable properties are often in tension: improving one behavior can impair another, so task progress can depend on alleviating the…
- arXiv cs.AIresearchCoverage Before Control: Route-Instruction Grounding and Steering for Controllable Retrosynthesis
arXiv:2609.39955v1 Announce Type: new Abstract: Single-step retrosynthesis models are commonly evaluated by their ability to recover recorded reactions. In practice, chemists may need to choose among several precursor se…
- arXiv cs.AIresearchPTNO: Training Neural Operators with Noisy Monte Carlo Estimates for Particle Transport Problems
arXiv:2609.40090v2 Announce Type: replace Abstract: Particle transport under multiple scattering is central to radiative transfer and plasma physics, yet high-fidelity Monte Carlo (MC) simulations must trace prohibitivel…
- arXiv cs.AIresearchUnlearnable, or Unmeasured? On the Reliability of Difficulty Labels in RLVR
arXiv:2609.40115v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has become an important approach for improving reasoning during post-training. Recent work suggests that some difficul…
- arXiv cs.AIresearchArgus: Academic Integrity in the Era of Generative AI
arXiv:2609.36073v1 Announce Type: cross Abstract: The rapid proliferation of large language models (LLMs) in the context of education has introduced significant challenges in enforcement of academic integrity, especially…
- arXiv cs.AIresearchTravel Time Prediction in Supply Chain Management Using Machine Learning
arXiv:2609.38190v1 Announce Type: cross Abstract: The purpose of this research is to find data and methods using machine learning and deep learning to correctly predict the estimated travel time for transportation and lo…
- arXiv cs.AIresearchA Moving-Horizon Approximate Branch-and-Reduce Method for Deep Classification Trees
arXiv:2609.38194v1 Announce Type: cross Abstract: Despite the importance for interpretability, decision trees face severe scalability challenges. Existing global optimal methods are often limited by binary feature select…