AI news archive.
10,661 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchPILLAR: Private Inverted-Index Lexical Lookup for Augmented Retrieval
arXiv:2609.36326v2 Announce Type: replace Abstract: Retrieval-augmented generation (RAG) hands the user's query to whoever hosts the corpus. We propose PILLAR, a Privacy-Preserving RAG (PPRAG) system based on Private Inf…
- arXiv cs.AIresearchThuRunel: Dynamic Decoupling for Structured Advisory Dialogue
arXiv:2609.36340v1 Announce Type: new Abstract: High-stakes advisory domains such as medical aesthetics, legal consultation, and educational planning exhibit a two-phase structure. The early phase requires empathetic eli…
- arXiv cs.AIresearchHyperZip: Efficient Data Compression through Personalized Diffusion LLMs with Hypernetworks
arXiv:2609.36357v1 Announce Type: new Abstract: Large language models (LLMs) have shown strong potential for lossless data compression, but existing approaches are constrained by the high computational cost and low throu…
- arXiv cs.AIresearchBetter Nearest Neighbor Graph Indices via (Efficient) LLM-Guided Pruning
arXiv:2609.36359v1 Announce Type: new Abstract: Graph-based approximate nearest neighbor search (ANNS) is widely used for large-scale semantic search. Its indices are constructed primarily based on geometric relationship…
- arXiv cs.AIresearchSupport-Set Target Leakage in Relational Foundation Models during In-Context Learning: Impact, Detection, and Mitigation
arXiv:2609.36384v1 Announce Type: new Abstract: Relational in-context learning (ICL) conditions predictions on the labeled support examples and their linked tables, creating a failure mode when the support set contains t…
- arXiv cs.AIresearchPersona Dosing: Calibrated Activation Steering for Graded Trait Control
arXiv:2609.36388v1 Announce Type: new Abstract: An activation-steering coefficient sets intervention strength, but requesting a particular degree of persona expression requires a behavioral scale. We study persona dosing…
- arXiv cs.AIresearchSupport-Set Target Leakage in Relational Foundation Models during In-Context Learning: Model Dependence and Evaluation Reliability
arXiv:2609.36417v1 Announce Type: new Abstract: Relational in-context learning (ICL) uses labeled support examples and their linked relational context to predict labels for new queries. This creates a failure mode when t…
- arXiv cs.AIresearchThe Safety Operator: Modulating the Expression of Safety Instructions via Spectral Optimization
arXiv:2609.36434v1 Announce Type: new Abstract: Context tokens in a transformer-based language model can be absorbed into the model's weights as a multiplicative operator. We study this operator in the setting of safety…
- arXiv cs.AIresearchBits Under ZK-LLM: Evaluating Zero-Knowledge-Friendly Quantization for Verifiable Private LLM Inference
arXiv:2609.36437v1 Announce Type: new Abstract: Zero-knowledge proofs are emerging as a promising approach for enabling private, verifiable LLM governance and auditing, where regulators, users, and auditors need to verif…
- arXiv cs.AIresearchRethinking Reasoning Paths as Phase-Structured Trajectories
arXiv:2609.36461v1 Announce Type: new Abstract: Large language models often improve problem-solving performance by generating multi-step reasoning paths, yet how to analyze the hidden states along these paths remains unc…
- arXiv cs.AIresearchGoing Beyond State-Reaching: Learning Abstractions for Intrinsically Motivated Option Discovery
arXiv:2609.36473v1 Announce Type: new Abstract: Temporal abstraction via options can improve exploration in large environments. However, existing option discovery algorithms find subgoals that target all aspects of the s…
- arXiv cs.AIresearchHuman-AI Collaboration: From Paradoxes to Patterns
arXiv:2609.36481v1 Announce Type: new Abstract: Evidence shows that humans and AI systems perform better together, by collaborating, than alone. This paper examines two key design dimensions of human-AI collaboration (au…
- arXiv cs.AIresearchAVIO: Learning to Add and Remove Sounding Objects in Audiovisual Scenes
arXiv:2609.36503v1 Announce Type: new Abstract: Adding or removing a sounding object requires coordinated changes to visual content and sound while preserving the surrounding scene. Yet paired supervision for localized n…
- arXiv cs.AIresearchVisual sensitivity is not claim retractability: persistence-aware credit assignment for multimodal reinforcement learning
arXiv:2609.36572v3 Announce Type: replace Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has been extended to Large Vision-Language Models (LVLMs), and perception-aware methods further encourage policies…
- arXiv cs.AIresearchMemEvo: Automatic Discovery of Streaming Video Memory Mechanisms
arXiv:2609.36581v1 Announce Type: new Abstract: Query-agnostic streaming video understanding requires vision-language models to continuously compress an indefinitely growing visual stream into a bounded memory before fut…
- arXiv cs.AIresearchTransformers Stop Thinking Too Early, and a Tiny LoRA Fixes It
arXiv:2609.36585v1 Announce Type: new Abstract: Pretrained transformers use little of their depth to follow references in context. Thirteen base models reliably follow only 1.4-3.6 lines, and extra pretrained loops add l…
- arXiv cs.AIresearchSAKI: Maximal-Coupling-Routed Teacher Supervision for On-Policy Distillation
arXiv:2609.36601v1 Announce Type: new Abstract: On-policy distillation (OPD) reduces train-test state mismatch by training a student on its own generated trajectories, but weak students may visit teacher-misaligned prefi…
- arXiv cs.AIresearchNeural Structural Reasoner: A Brain-inspired Architecture for Reasoning over Structured Knowledge
arXiv:2609.36620v1 Announce Type: new Abstract: Structural reasoning, the ability to recognize and make inferences over the relational structure between objects and concepts, is a hallmark of human cognition, yet prevail…
- arXiv cs.AIresearchDualTrack: Synchronized speech-gesture generation via symmetric coupling of pretrained priors
arXiv:2609.36624v1 Announce Type: new Abstract: Joint speech-gesture synthesis must coordinate two modalities despite limited paired data. Existing approaches often lack bidirectional interaction, have limited language c…
- arXiv cs.AIresearchRankBuffer: Efficient Ranking-Based Rewards for Open-Ended Generation
arXiv:2609.36652v1 Announce Type: new Abstract: Open-ended generation lacks canonical answers, making pointwise rewards difficult to calibrate for group-based reinforcement learning. Directly ranking same-query rollouts…
- arXiv cs.AIresearchFineSID: Scalable and Efficient Semantic Identifier Learning for Generative Recommendation
arXiv:2609.36670v2 Announce Type: replace Abstract: A critical prerequisite of generative recommendation is designing semantic identifiers (SIDs) that are both scalable to large item sets and efficiently learnable. Exist…
- arXiv cs.AIresearchFairDiff: Mitigating the Self-Reinforcing Matthew Effect in Diffusion Recommender Models
arXiv:2609.36671v2 Announce Type: replace Abstract: While the "Matthew Effect" and filter bubbles are widely recognized outcome-level biases in recommender systems, we reveal that Diffusion Recommender Models (DRMs) uniq…
- arXiv cs.AIresearchJudgeProfile: Understanding and Steering Subjectivity in LLM Judges
arXiv:2609.36705v2 Announce Type: replace Abstract: LLM judges are inherently subjective, often favoring different responses in pairwise comparison when neither option is objectively wrong. To study this subjectivity, we…
- arXiv cs.AIresearchBiFE: Search-Efficient Discovery of CPU-Only Branching Policies via LLM-based Bi-Fidelity Evolution
arXiv:2609.36735v1 Announce Type: new Abstract: In branch-and-bound (B&B) for mixed-integer linear programming (MILP), branching variable selection critically impacts efficiency. Existing neural branching policies often…
- arXiv cs.AIresearchSIPO: Unifying Reinforcement Learning with On-Policy Self-Distillation
arXiv:2609.36742v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for improving large language models (LLMs) on various tasks, yet its sparse outcome rew…
- arXiv cs.AIresearchGeneralizable Lifelong Model Editing via Preference Optimization
arXiv:2609.36748v1 Announce Type: new Abstract: Knowledge editing enables rapid updates of specific factual knowledge in large language models (LLMs) without full retraining. However, more realistic scenarios call for a…
- arXiv cs.AIresearchAperture: Merge-Consistent Rotary States for Compressed Tokens
arXiv:2609.36781v1 Announce Type: new Abstract: Token compression combines content from several positions, yet rotary position embeddings usually assign the merged token one coordinate. We ask what positional information…
- arXiv cs.AIresearchAI as a Compiler: Compiling Triton kernels without the Triton compiler
arXiv:2609.36800v1 Announce Type: new Abstract: Compiler backends are expensive to build and maintain as programming models, workloads, and accelerators evolve. We investigate whether large language models can replace th…
- arXiv cs.AIresearchCANTO: CAD-Native Transformer Operators for AI-Aided Engineering
arXiv:2609.36806v2 Announce Type: replace Abstract: Modern engineering systems, from automobiles to aircraft, are designed by using precise, continuous parametric computer-aided design (CAD) models. Evaluating design cha…
- arXiv cs.AIresearchGeometry-Conditioned Fixed-Scaffold Encoders for Time-Warp Robust Sequence Retrieval
arXiv:2609.36809v1 Announce Type: new Abstract: Embedding-based retrieval is attractive for long sequence collections because each item can be encoded once and searched by nearest-neighbor ranking. The difficulty is that…
- arXiv cs.AIresearchCalibrate the Decisions That Change the Future: On-Policy Post-Training Quantization for Multimodal Large Language Models
arXiv:2609.36828v1 Announce Type: new Abstract: Post-training quantization (PTQ) lowers deployment cost for multimodal large language models, but calibration typically reconstructs fixed sequences with local objectives.…
- arXiv cs.AIresearchWhere Does Staleness Accumulate? Pool Aware Effective Staleness Control for Asynchronous RL in LLM Post-Training
arXiv:2609.36830v1 Announce Type: new Abstract: Fully asynchronous reinforcement learning (RL) improves resource utilization in large language model post-training by overlapping rollout generation with policy optimizatio…
- arXiv cs.AIresearchARC-KV: Amortizing Anchor Search for Reconstruction-Based KV Cache Compaction
arXiv:2609.36835v1 Announce Type: new Abstract: Long-context large language model inference is bottlenecked by KV caches that grow linearly with sequence length. This burden is especially severe for long, reusable contex…
- arXiv cs.AIresearchAutomated Screw Planning for Reduced Pelvic Fractures Based on Statistical Shape Models and Deep Learning
arXiv:2609.36847v1 Announce Type: new Abstract: Percutaneous iliosacral screw fixation is an important minimally invasive treatment for unstable pelvic fractures. Because the sacroiliac region has complex anatomy and nar…
- arXiv cs.AIresearchIronLLM: Forging Compact Edge-Native Language Models for Real-Time Embodied Intelligence
arXiv:2609.36860v1 Announce Type: new Abstract: We present IronLLM-0.6B, a 654M-parameter language model designed for efficient on-device inference. IronLLM-0.6B combines a hybrid attention architecture with X-MTP, a lig…
- arXiv cs.AIresearchBeyond Sub-Gaussian Detector Scores: Robust Weighted Profile-Loss Change Point Detection for Human-LLM Text Segmentation
arXiv:2609.36888v1 Announce Type: new Abstract: Mixed human-LLM documents require locating authorship transitions from detector scores whose reliability varies across text units. Existing weighted mean contrasts are vuln…
- arXiv cs.AIresearchHorizonFlow: Variable-Length Planning for Offline Goal-Conditioned RL
arXiv:2609.36896v1 Announce Type: new Abstract: Recent advances in generative planning have made trajectory inpainting a promising approach to offline goal-conditioned reinforcement learning. However, these methods typic…
- arXiv cs.AIresearchSTAR-GRPO: Canonical Anchoring and Reliability-First Advantages against Representation-Dependent Reward Hacking
arXiv:2609.36900v1 Announce Type: new Abstract: Reward hacking occurs when policy optimization exploits a brittle reward interface or an overly permissive proxy objective, improving the training score without improving t…
- arXiv cs.AIresearchLearn from the Gap: Differential-Aware Advantage Pruning with Adaptive Rollout Sampling for GRPO
arXiv:2609.36932v3 Announce Type: replace Abstract: Recently, Group Relative Policy Optimization (GRPO) and its variants have been developed for policy optimization and demonstrated notable performance gains. However, th…
- arXiv cs.AIresearchVLALight: A Vision-Language-Action Model for Traffic Signal Control
arXiv:2609.36934v1 Announce Type: new Abstract: Traffic signal control (TSC) is essential for improving urban mobility and reducing congestion. Although roadside cameras are widely deployed at signalized intersections an…
- arXiv cs.AIresearchCoEM: Empowering Long-Context Reasoning with Commit-on-Evidence Memory
arXiv:2609.36935v2 Announce Type: replace Abstract: Long-context reasoning is essential for complex and long-horizon tasks, yet the performance of large language models (LLMs) degrades as context length increases. Recent…
- arXiv cs.AIresearchDual-Channel Robust Group-Relative Policy Optimization via Advantage and Sequence-Weight Estimation
arXiv:2609.36944v1 Announce Type: new Abstract: Group-relative policy optimization relies on reward-derived advantages and sequence-level likelihood weights, both of which can be sensitive to localized outliers. Extreme…
- arXiv cs.AIresearchREALHOP: Rethinking Multi-Hop Reasoning Evaluation via Behavioral Auditing
arXiv:2609.36984v1 Announce Type: new Abstract: Complex questions often require multi-hop reasoning that connects facts distributed across sources or distant regions of a long context through intermediate steps. Benchmar…
- arXiv cs.AIresearchCF-LoRA: Decoupled Factor Aggregation and Adaptation-Aware Client Clustering for Federated LoRA Fine-Tuning
arXiv:2609.36986v1 Announce Type: new Abstract: Federated LoRA fine-tuning enables parameter-efficient adaptation of pre-trained models without sharing private data, but suffers from two fundamental mismatches under hete…
- arXiv cs.AIresearchImbalancE: Inference-Time Latent Search Against Degree Imbalance in Link Prediction
arXiv:2609.36996v1 Announce Type: new Abstract: Knowledge Graph Embedding models have been extensively used to learn representations of entities and relations in Knowledge Graphs for predicting missing links. However, th…
- arXiv cs.AIresearchLanguage as the Interface: Foundation-Model Contrastive Learning Links Transcriptomes and Electrophysiology
arXiv:2609.37024v1 Announce Type: new Abstract: Integrating transcriptomic and electrophysiological data is essential for building multimodal foundation models for neuroscience. Patch-seq provides paired measurements of…
- arXiv cs.AIresearchBeyond Low-Rank Parameterization: Narrowing the Gap Between LoRA and Full Fine-Tuning via Gradient Decomposition
arXiv:2609.37027v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) is a widely used approach to parameter-efficient fine-tuning (PEFT), yet a performance gap can remain relative to full fine-tuning (FFT). Many Lo…
- arXiv cs.AIresearchFedLAFP: Low-Rank Aggregation Meets Full-Rank Personalization in Federated Fine-Tuning
arXiv:2609.37033v1 Announce Type: new Abstract: Federated parameter-efficient fine-tuning enables clients to adapt pre-trained models without sharing raw data or communicating the full model, but statistical heterogeneit…
- arXiv cs.AIresearchWatch-Think-Interact: Bootstrapping Long-Horizon Multi-Turn Streaming Video Reasoning with Reinforcement Learning
arXiv:2609.37035v1 Announce Type: new Abstract: Streaming video assistance requires models to answer asynchronous questions from an observed prefix under a fixed context budget. Existing approaches model response timing…
- arXiv cs.AIresearchACTR: Aligning Thoughts and Responses for Multilingual Safety in Reasoning LLMs
arXiv:2609.37054v2 Announce Type: replace Abstract: Ensuring the safety of reasoning large language models (LLMs) across languages is essential for their reliable deployment. However, when exposed to jailbreak attacks in…