arXiv AI digest

53 days

October 6, 2026Tuesday
1. Towards Looped Models Done Right, Part II: Rethinking at Fixed Points
2. MemPilot: Orchestrating On-Demand Multimodal Memory Curation for LLM Agents
3. CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling
4. T-Search: An Open Agentic Retriever and Playground for Hard Multi-Step Search
5. One Global Beam Across Many GPUs: High-Throughput Beam Search at Billion-Record Frontier Scale
cs.LGcs.CLcs.DC
October 5, 2026Monday
1. FrugalEvo: Towards Cost-Aware LLM-Guided Program Evolution
2. Credit Where It Matters: Dependency-Aware Policy Optimization for Terminal Agents
3. NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents
4. MRVQ: One Resident Index for Dimension- and Rate-Elastic Vector Search
5. Rethinking What to Cache in Few-Step Diffusion Transformers: Solver-Aware Target Selection
cs.NEcs.AIcs.CV
October 4, 2026Sunday
1. Trust the Direction, Search the Step: Zero-and-First-Order Methods for LLM Fine-Tuning
2. Decoding Looped Transformers Better for (Almost) Free
3. From Knowledge Access to Source Learning: Developing Source-Specific Competence
4. Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes
5. LLM2Jev: LLMs Are Already Jev-Style Decision Models -- When and How to Fine-Tune Them
cs.LGcs.CLcs.CV
October 3, 2026Saturday
1. KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards
2. TACO: Ternary Absolute-max Column-wise One-sparse Optimizer for LLM Fine-Tuning
3. AutoCompact: Learning When to Compact Context in Long-Horizon Coding Agents
4. Keyword Harnesses Fail Open: A Cheap Diagnostic Ladder for Tool-Use Claims in Small Language Models
5. Argo-Bench: Evaluating Data Agents on Enterprise-Scale Workflows
cs.CLcs.LG
October 2, 2026Friday
October 1, 2026Thursday
1. Turbo Harness: Instance-Adaptive Harness Optimization
2. Compression Footprints as Security Signals for Model-Poisoning Defense in Federated Learning
3. cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents
4. ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents
5. PhantomEnvironments: Training LLM Agents in Fictional Worlds
cs.AIcs.LGcs.CV
September 30, 2026Wednesday
1. WUSH-KV: KV Cache Quantization with Data-Adaptive Transforms
2. Mira: Memory-Efficient MoE Inference Using Adaptive Caching and Predictive Expert Staging
3. Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution
4. Thinking Before Thinking: Scaling Agentic Inference Through Meta-Reasoning
5. AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation
cs.LGcs.AI
September 29, 2026Tuesday
1. TokenCast: Forecasting Token Consumption During LLM Agent Execution
2. KV-streams for Efficient Compaction in Agentic Reinforcement Learning
3. Failure-Transparent Agents: Benchmarking Post-Failure Reporting in Tool-Using Language Models
4. Report: Progressive Disclosure of Agent Skills
5. PhoneCLI: From App Interfaces to Callable Commands for Mobile Agents
cs.LGcs.AI
September 28, 2026Monday
September 27, 2026Sunday
September 26, 2026Saturday
September 25, 2026Friday
September 24, 2026Thursday
September 23, 2026Wednesday
1. SWE-Serve: Benchmarking Agentic Engineering For Production Inference Serving
2. SARA: SLO-Aware Resource Allocation for Disaggregated Agentic LLM Services
3. CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents
4. Measuring the Serving Stack Instead of the Model: Hidden Confounds in Local Tool-Use Evaluation
5. Greedy Decoding Is Not Precision-Invariant: Cross-Precision Output Divergence in LLM Inference
cs.AIcs.PFcs.CLcs.LG
September 22, 2026Tuesday
1. Who Pays for the KV Cache? Attributing Shared AI Inference Spend Across Kubernetes and LLM Provider Bills
2. Harness-Zero: Harness Distillation via Agent-as-Harness
3. Trajectory-Aware Benchmark Subset Selection for Cost-Efficient Software Engineering Agent Regression Testing
4. OSWorld-Pro: Process-based Evaluation for Computer Use Agents
5. SPECTRA: Adaptive Execution of Speculative Decoding on a Runtime-Reconfigurable Tiled Architecture
cs.DCcs.AIcs.SEcs.CLcs.AR
September 21, 2026Monday
1. Gricea: An Open Science Platform for Conversational AI Research
2. RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents
3. NemotronLabs VoiceChat: An Open Full-duplex Speech-to-Speech Model with Tool Calling Capabilities
4. How Researchers Use and Verify AI Coding Assistants: Tasks and Validation Practices in Scientific Programming
5. An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency
cs.HCcs.CLcs.SE
September 20, 2026Sunday
1. Chronicle: Cut-Point Replay for Regression Testing of LLM Agents
2. Inference-Engine Fingerprinting Attacks are Practical: Exploring Model-Driven Environmental Discovery, Exploitation, and Escape
3. COIN-GP: Cooperative Online Learning in Networked Distributed Systems with Partial Measurements via Gaussian Process Regression
4. Q&A on Any Spreadsheet Requires Interpreting Its Grid Structure
5. A Simulation Platform for AUV Fault Recovery: Exploring LLM-Based Diagnostic Strategies
cs.CLcs.CRcs.LGcs.AIcs.RO
September 19, 2026Saturday
1. dQwen3.5: Hybrid-Attention Diffusion Language Models
2. On-Demand Attention: Language Models Know When to Recall
3. Deep Noir: Autonomous Steering Discovery via Architectural Chronometry in Transformer Models
4. Multi-center Medical Data Mining with FL-Net - A One-stop Shop for Federated Learning
5. HIL-UMI: Bringing Human-in-the-Loop Post-Training of Vision-Language-Action Models to Universal Manipulation Interface
cs.CLcs.AIcs.LGcs.RO
September 18, 2026Friday
1. An Empirical Study of Harness Design for Coding Agents
2. RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents
3. PixelFlow: Token-Level Workload Management for Efficient Distributed DiT Serving
4. Quantifying Overclaiming Propensity in Frontier LLM Agents
5. Large Language Models as Falsifiers for Cyber-Physical Systems
cs.AIcs.DCcs.SEeess.SY
September 17, 2026Thursday
1. Replication-Aware Placement of Functions and Data in the Edge-Cloud Continuum
2. Ermes: a Stateful Serverless Platform for the Edge-to-Cloud Continuum
3. Cognitive Extensions for Dual-Process Language Agents: Memory and Self-Reflection in Interactive Environments
4. Affora: A Design System for Agent-Friendly Interfaces
5. Taming the Agentic RAN: Stability-Guaranteed Arbitration of Autonomous AI Agents in O-RAN
cs.DCcs.AIcs.HCcs.NI
September 16, 2026Wednesday
1. JustFit: 200K-Token LLM Serving on a 24 GiB Laptop with Just-in-Time State Management
2. FlashVector: Agent for Hierarchical Model Serving Stack Optimization
3. Coding Agents Have Converged: Why the SWE-bench Leaderboard Can No Longer Order Its Top Entries, and What to Measure Instead
4. ScienceBuddy: Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents
5. Mo' Models, Mo' Problems: How to best select model pools when designing Multi-Agent Systems
cs.AIcs.SEcs.MA
September 15, 2026Tuesday
September 14, 2026Monday
September 13, 2026Sunday
September 12, 2026Saturday
1. CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search
2. AdamX: Cosine similarity meets gradient descent
3. RetroThinker: Enabling Retrospective Thinking in Speech LLMs
4. Predicting Privacy Leakage from Weight Spectral Density
5. RAG-Safety-Bench: Reliable Evaluation of Retrieval-Augmented LLM Safety
cs.LGeess.AScs.CL
September 11, 2026Friday
1. GPU-CFR: 80x Faster Counterfactual Regret Minimization by Compiling the Game to Static Dataflow and CUDA Graph Replay
2. Can Edge-Deployable Vision-Language Models Identify Species?
3. Domain-Specific Hallucination Detection in Large Language Models
4. SpecGuard: Inference-Time Backdoor Detection For Free
5. An analysis of the relationship of input metrics
cs.DCcs.AIcs.CLcs.CRcs.SE
September 10, 2026Thursday
1. Avatar: Toward Autonomous End-to-End Orchestration of Scientific Workflows using LLMs
2. IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model Identifier
3. TrajMark: Ownership Attribution and Segment-Level Tamper Localization for Coding-Agent Trajectories
4. Towards Scalable and Cost-Efficient Vulnerability Detection: A Study on Automatic Query Generation
5. dexamine: A Python package for Uniswap event data on Ethereum
cs.DCcs.CLcs.CRcs.SEq-fin.TR
September 9, 2026Wednesday
1. Distributed Linear Programming on GPU Clusters at Extreme Scale
2. ExecCritic: Learn to Test, Test to Improve for Coding Agents
3. ToolLoop: Closed-Loop Tool-Use Data Synthesis via Decomposed Generation and Dynamic Self-Feedback
4. Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails
5. PlayTrain: An Efficient Reinforcement Learning Framework for LLM-Generated Adaptable JavaScript Games
math.OCcs.AIcs.CLcs.LG
September 8, 2026Tuesday
1. Design Docs Are All You Need: An AI-native Machine-Learning Performance Tool
2. Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory Portability
3. The History Is the Detector: Executing CVE Patch History, End-to-End
4. CONTINUITY: Security-Context Contracts for Composable LLM Agent Controls
5. Trace2Tower: Transition-Aware EigenTrace Induction of Multi-Level Skills for LLM Agents
cs.PLcs.AIcs.CR
September 7, 2026Monday
1. Beyond Scalar Flexibility: From Eligible AI Workloads to Dependable Load Relief
2. Multi-Step Tool-Calling over Korean Open Public APIs: A Benchmark and a Data-Synthesis Recipe
3. CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents
4. Testing Interchangeability in LLM Agent Teams
5. How to Speculate about Uncertainty in Agentic Coding? A Draft-Model Gate Method
eess.SYcs.AIcs.LG
September 6, 2026Sunday
1. Clean Engineering, Unstable Measurement: A Preregistered Reliability Failure of Black-Box LLM Observers on Shared Endpoints
2. Terminal-Universe: Turning Agent Trajectories into Scalable Terminal Environments
3. Hardware-Aware FP4 FlashAttention-4
4. Why Gated DeltaNet Survives 4-Bit Quantization: NVFP4 W4A4 for the Recurrent Half of a Hybrid 27B LLM
5. Instruction Duplication as an Inference-Time Control Primitive
cs.AIcs.LG
September 5, 2026Saturday
1. Para-Pipe: Exploiting Hierarchical Operator Parallelism of ML Computational Graphs on SoCs
2. SENTINEL-RL: Offloading Topological Reasoning from LLM Agents in the Security Operations Center
3. A Low-Cost, Open Platform for End-to-End Autonomous Driving on a Miniature Ackermann Vehicle
4. Environment Evolution for Terminal Agents
5. The Dice Roll Method: A Standardized Protocol for Repeated-Query Auditing of Large Language Model Brand Recommendations
cs.DCcs.CRcs.LGcs.AIcs.IR
September 4, 2026Friday
1. SWE-Gate: Passing Functional Tests Is Not Enough for Software Engineering Agents
2. The Natural Language Interaction Protocol and Standard for AI Agents
3. PatchBench: Evaluating AI Agents for Vulnerability Patching
4. When Models Edit Too Much: On the Fidelity of Minimal Code Edits
5. LabelMate: An LLM-Driven Framework for Refined Issue Report Labeling
cs.SEcs.AIcs.CR
September 3, 2026Thursday
1. Type Hints in Python Libraries and Frameworks: An Empirical Analysis of Adoption and Maintenance
2. The Import Tax: A Longitudinal Measurement of Startup Cost in the Python Ecosystem
3. Measurement-Driven Sub-Network Selection for On-Premise Retrieval-Augmented Factory Agents
4. EarlyEval: Cheaper Agent Evaluation via Early Outcome Prediction
5. SafeEvolve: Harness-Policy Co-Evolution from Agent Experience for Safety Alignment
cs.SEcs.AIcs.CL
September 2, 2026Wednesday
1. From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix
2. Adaptive Critical Token-Aware Retrieval for Repository-Level Code Generation
3. Efficient SWE Agent Benchmarking via Trajectory-Aware Evaluation
4. When Guardrails Look Effective: Construct Validity Failures in LLM Agent Commerce Evaluation
5. Can LLMs Design Video Coding Tools? A Case Study on Planar Mode
cs.CLcs.SEcs.AIcs.MM
September 1, 2026Tuesday
1. LLM Post-Training as Brownfield Maintenance: An Industrial Perspective on Dataware Engineering
2. Token-Efficient Data Reasoning Agents via Adaptive Structuring of Unstructured Data
3. BLOOM-WILT: Logit Tilting for Behaviour Elicitation in Automated LLM Auditing
4. DIASENTINEL: An Auditable Multi-Agent System for Guideline-Grounded Diabetes Risk Screening
5. Context-Aware Interleaved Batching for WhisperX
cs.SEcs.AIcs.CL
August 31, 2026Monday
1. On the Maintenance and Co-evolution of Agent Plugins: An Empirical Study of Claude Code Plugin Marketplaces
2. LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment
3. Rethinking Vulnerability Remediation as a Capacity Allocation Problem
4. Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
5. Logos: An Agent Harness on a Cross-Process Bus
cs.SEcs.CRcs.AI
August 30, 2026Sunday
1. CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes
2. Beyond F1: Evaluating Coverage and Failure Recovery in AI Model Security Scanners
3. Token-Level Advertising
4. INTENT-AS-A-TOOL Makes it Easy to Track Agentic Misalignment
5. Decoupled I/O-Dominant Pipelines for Large-Scale Whole-Slide Image Embedding Extraction
cs.CLcs.CRcs.GTcs.DC
August 29, 2026Saturday
1. From Static to Dynamic: Benchmarking Real-World Code Review with MCR-Bench
2. RedEvoAgent: Automatic Red-Teaming Agent with Experience-Driven Skill Evolution
3. Scaling Graph Neural Networks for Friend Recommendation: Multi-Hash User Embeddings and Temporal Neighbor Sampling
4. Verify Smarter, Evolve Further: Efficient Harness Evolution through Behavior-Aware Verification
5. When Context Gets Root: Privilege Escalation in LLM Harnesses
cs.SEcs.CRcs.IRcs.AI
August 28, 2026Friday
1. WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution
2. SWE-Prime: Fewer Trajectories, Better Performance
3. Persona-Execution Separation: An Architecture Pattern for Evolving LLM Agents under Execution Audit
4. Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090
5. BTS-AgentBench: A Deterministic, Replayable Pipeline from Read-Only Telemetry Logs to Agent Benchmarks
cs.AIcs.SEcs.CL
August 27, 2026Thursday
1. Slasher: Power Flexibility for Cloud Datacenters
2. Trace Integrity for LLM Data Agents: A Vision for Auditable Structured Reasoning in Real-World Systems
3. ProgRouter: Online Progress-Guided Orchestration for Multi-Agent LLM Workflows under Quality-Cost Tradeoffs
4. A Self-Evolving Multi-Agent Framework Defense against LLM Jailbreak Attacks
5. TraceML: An Empirical Analysis of Human-Agent Planning in Machine Learning Development
cs.DCcs.AIcs.CRcs.LG
August 26, 2026Wednesday
1. Next-generation O-RAN Edge: Energy-aware Joint Placement and Migration of Cloud-Native Functions
2. BrowserForge: Scaling Web Episode via Parallel Browser Sandboxes
3. StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments
4. StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing
5. Prompt Structure Redistributes, Not Reduces: An Empirical Analysis of Security-Weaknesses in LLM-Generated Python Code
cs.ITcs.CLcs.AIcs.CR
August 25, 2026Tuesday
1. SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?
2. Prime Agent: A Self-Improving RLM Harness
3. Robustness of Anomaly Detection Models for Industrial Control Systems under Training-Time Data Contamination
4. InjecMEM: Memory Injection Attack on LLM Agent Memory Systems
5. RAD: Rule-Augmented Relational Anomaly Detection
cs.CLcs.AIcs.CRcs.LG
August 24, 2026Monday
1. Natural-Language Workflows Are Not Software Yet: Artifact-Driven Compilation for Reliable Agent Execution
2. AI-to-AI Code Reviews of GitHub Pull Requests
3. Beyond Fault Localization: A Trajectory-Level Study of LLM Agents for Microservice Root Cause Analysis
4. Utility Under Attack: Agent Memory Poisoning and the Limits of Content Screening and Provenance Ranking
5. Ontology-supported AI Model and Dataset Management
cs.SEcs.CRcs.AI
August 23, 2026Sunday
1. An Agentic Approach for Active Data Collection, Travel Behavior Modeling, and Weather-Sensitive Demand Prediction
2. Inducing Task Models from Computer-Use Traces
3. AI4AI-Bench: Benchmarking LLM Agents in Algorithmic Design for Recursive Self-Improvement
4. Phantom Gains: Auditing Self-Improvement Against a Measured Null
5. Break It Down, Pass It On: Cross-Task Skill Transfer in LLM Agents
cs.AIcs.CL
August 22, 2026Saturday
1. Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization
2. BreakGuard: Towards Detecting Dependency Breaking Changes with LLM-Generated Tests
3. Task-CoEvolve: Efficient Harness Optimization via Adaptive Validation Task Selection
4. From Agent Behaviour to Agent-Friendly Documentation: An Empirical Study of How Coding Agents Discover, Read, and Write Technical Documentation
5. Learning When to Think: Adaptive Reasoning for Test-Time Compute Allocation
cs.CLcs.SEcs.AI
August 21, 2026Friday
1. Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation
2. Design and Empirical Evaluation of a Network-Centric, On-Premises Architecture for Earth Observation Data Access
3. Which Eviction Policy Should an LLM Cache Use? A Systematic Study Across Workloads, Capacities, and Encoders
4. Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference
5. MidTool: Mid-training Data Synthesis for Agentic Tool Use
cs.AIcs.DCcs.DBcs.IR
August 20, 2026Thursday
1. SPADE: Self-Play in Adaptive Synthetic Executable Environments
2. Beyond the Transcript: Detecting Covert Co ordination in Latent Multi-Agent Communication
3. Pre-Compiled Pipeline Shards for Distributed LLM Inference on Intel AI PC Fleets
4. Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation
5. What is Missing from AI Post-Training AI: An Empirical Analysis
cs.CLcs.AIcs.DCcs.LG
August 19, 2026Wednesday
1. StagedWorkspace: A Versioned Workspace for Knowledge-Work Agents
2. The Polyglot's Dilemma: Conformance Testing a Dozen Specs in as Many Languages
3. Minimizing Commit Rules for DAG-based Atomic Broadcast
4. Too Sure to Be Safe: Model Calibration for Reliable Log Anomaly Detection
5. Judge, Retrieve, or Abstain: Uncertainty-Guarded LLM Judging with Provable Risk Guarantees
cs.AIcs.SEcs.DCcs.LGcs.CL
August 18, 2026Tuesday
1. zLend: A Dual-Scope Cash-Flow Reconstruction Framework for On-Chain Credit Underwriting
2. Model Hypnosis: Strong control of AI via additive subliminal effects
3. GEO-Flag: Detecting and Measuring GEO-Optimized Web Content
4. Quipu: A Governed Bitemporal Knowledge Graph Store
5. ClawGym II: Exploring Black-Box RL on Agent Harness
6. LAVA: Logic-Aware Validation and Augmentation Framework for Large-Scale Financial Document Auditing
7. Le Critique: Privileged Value Functions for LLM Reinforcement Learning
8. Semantic Bandits: In-Context Exploration-Exploitation is Biased by Semantic Priors
9. UniTAC: Universal Task-Aware Compression via Weighted Distortion Measures
10. Steering the Flow: Inverting Face Recognition Models via Gradient-Guided Flow Matching
q-fin.RMcs.CLcs.LGcs.AIcs.CV
August 17, 2026Monday
1. Handover of In-Context Learning State Across Session Boundaries
2. Split the Labor: Separating Evidence Interpretation from Decision Aggregation
3. Twin: Playing an Unknown Game with a Test-Time Digital Twin
4. You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model
5. SheetCompass: Hierarchical Relation Graphs for Agentic Spreadsheet Reasoning
6. Wyvern: An Agentic Framework for Generating Grounded Multimodal Reports
7. PACE-Bench: Benchmarking Physics Adaptation via Code Evolution in Dynamic Environments
8. Knowing When to Stop: Bayesian Optimal Stopping for LLM Evaluations
9. DeaMoE: Efficient MoE Structure for Fast Small-Batch Decoding
cs.AIcs.CLcs.LG
August 16, 2026Sunday
1. LittleLearner: Language Models Under Pedagogically Controlled Knowledge Exposure
2. Reduced Matrix Multiplication: Input-Adaptive Matrix-Product Reduction for LLM Inference
3. MLLM-Routed Heterogeneous Ensembles for Robust Cross-Dataset Image Classification
4. OmniScientist: An Omni-Modal Omni-Discipline AI Scientist
5. SAEVerbalizer: Generating Explanations for Sparse Autoencoder Features via Representation Verbalization
6. Intern-S2-Preview: Scientific Agentic Foundation Model
cs.CLcs.LGcs.CVcs.AI
August 15, 2026Saturday
1. MARC v1: An Open-Source Multi-Agent Framework for Clinical AI Reasoning and Coordination
2. TraVEL: Trajectory-Guided Video Embedding Learning for Driving-Video Retrieval
3. DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data
cs.AIcs.CVcs.CL