Papers

New on arXiv

Sun 5 Apr 2026

196 papers

meta-pipe: An LLM-agent pipeline for end-to-end automated systematic review and meta-analysisLLM agents and tool use → Web, search, and research agents
Your Agent Has a Genome: Sequence-Level Behavioral Analysis and Runtime Governance of LLM-Powered Autonomous AgentsLLM agents and tool use → Software engineering coding agents
StateSMix: Online Lossless Compression via Mamba State Space Models and Sparse N-gram Context MixingState-space sequence models (S4, Mamba) → Core SSM architectures for language
AI Agents for Sustainable SMEs: A Green ESG Assessment FrameworkLLM agents and tool use → General LLM agent frameworks and benchmarks
Belief-Aware VLM Model for Human-like ReasoningLLM reasoning → Multimodal reasoning
ZitPit: Consumer-Side Admission Control for Agentic Software IntakeLLM agents and tool use → Software engineering coding agents
The Art of Building Verifiers for Computer Use AgentsLLM agents and tool use → GUI and mobile agents
FreakOut-LLM: The Effect of Emotional Stimuli on Safety AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Architecture Without Architects: How AI Coding Agents Shape Software ArchitectureLLM agents and tool use → Software engineering coding agents
SkillAttack: Automated Red Teaming of Agent Skills through Attack Path RefinementLLM agents and tool use → Agent threats, guardrails and governance +1
Cactus: Accelerating Auto-Regressive Decoding with Constrained Acceptance Speculative SamplingEfficient LLM inference and serving → Speculative decoding
Enhancing sample efficiency in reinforcement-learning-based flow control: replacing the critic with an adaptive reduced-order modelDeep reinforcement learning methods → RL with diffusion and flow models
High-Stakes Personalization: Rethinking LLM Customization for Individual Investor Decision-MakingLLM reasoning → Reasoning for agents, planning and code
Correcting Source Mismatch in Flow Matching with Radial-Angular TransportDiffusion models → Flow matching, time series and science
InferenceEvolve: Towards Automated Causal Effect Estimators through Self-Evolving AISynthetic data and self-improvement for LLMs → Self-evolving LLM agents
Beyond Fluency: Toward Reliable Trajectories in Agentic IRLLM agents and tool use → Software engineering coding agents
Governance-Constrained Agentic AI: Blockchain-Enforced Human Oversight for Safety-Critical Wildfire MonitoringLLM agents and tool use → Agent threats, guardrails and governance
APPA: Adaptive Preference Pluralistic Alignment for Fair Federated RLHF of LLMsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Towards Unveiling Vulnerabilities of Large Reasoning Models in Machine UnlearningLLM reasoning → Reasoning for agents, planning and code +1
MC-CPO: Mastery-Conditioned Constrained Policy Optimization for Pedagogically Safe Intelligent Tutoring SystemsDeep reinforcement learning methods → RL training of language and agent models
Combee: Scaling Prompt Learning for Self-Improving Language Model AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks +1
Pedagogical Safety in Educational Reinforcement Learning: Formalizing and Detecting Reward Hacking in AI Tutoring SystemsLLM alignment / RLHF / preference optimisation → Reward hacking and misspecification +1
Subspace Control: Turning Constrained Model Steering into Controllable Spectral OptimizationLLM safety and security → Machine unlearning
Three Phases of Expert Routing: How Load Balance Evolves During Mixture-of-Experts TrainingMixture-of-experts → Expert routing strategies
SW-A2-Bench: Benchmarking Autonomous Software Agent Generation for Agentic WebLLM agents and tool use → Software engineering coding agents
Learning from Imperfect Demonstrations via Temporal Behavior Tree-Guided Trajectory RepairDeep reinforcement learning methods → RL for robot locomotion and manipulation
TimeSeek: Temporal Reliability of Agentic ForecastersLLM agents and tool use → Financial trading and forecasting agents
Sharp asymptotic theory for Q-learning with LDTZ learning rate and its generalizationDeep reinforcement learning methods → RL theory and sample complexity
DARE: Diffusion Large Language Models Alignment and Reinforcement ExecutorLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
LOCARD: An Agentic Framework for Blockchain ForensicsLLM agents and tool use → Agent threats, guardrails and governance
Don't Blink: Evidence Collapse during Multimodal ReasoningLLM reasoning → Multimodal reasoning
ClawArena: Benchmarking AI Agents in Evolving Information EnvironmentsLLM agents and tool use → General LLM agent frameworks and benchmarks
DriveVA: Video Action Models are Zero-Shot DriversWorld models → World models for autonomous driving
GENFIG1: Visual Summaries of Scholarly Work as a Challenge for Vision-Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Incomplete Multi-View Multi-Label Classification via Shared Codebook and Fused-Teacher Self-DistillationKnowledge distillation → General knowledge distillation methods
Primitive-based Truncated Diffusion for Efficient Trajectory Generation of Differential Drive Mobile ManipulatorsDiffusion models → Diffusion and flow policies for robotics
Adaptive Action Chunking at Inference-time for Vision-Language-Action ModelsVision-language-action models → Efficient VLA inference and adaptation
Readable Minds: Emergent Theory-of-Mind-Like Behavior in LLM Poker AgentsLLM agents and tool use → Reinforcement learning for LLM agents
The Geometric Alignment Tax: Tokenization vs. Continuous Geometry in Scientific Foundation ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Many Preferences, Few Policies: Towards Scalable Language Model PersonalizationLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
OP-GRPO: Efficient Off-Policy GRPO for Flow-Matching ModelsDiffusion models → Diffusion and flow policies for robotics
Profile-Then-Reason: Bounded Semantic Complexity for Tool-Augmented Language AgentsLLM reasoning → Reasoning for agents, planning and code +1
Shorter, but Still Trustworthy? An Empirical Study of Chain-of-Thought CompressionLLM reasoning → Test-time compute and efficient thinking
InsTraj: Instructing Diffusion Models with Travel Intentions to Generate Real-world TrajectoriesDiffusion models → Diffusion and flow policies for robotics
Restless Bandits with Individual Penalty Constraints: Near-Optimal Indices and Deep Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
Fine-grained Analysis of Stability and Generalization for Stochastic Bilevel OptimizationDeep reinforcement learning methods → RL theory and sample complexity
CoopGuard: Stateful Cooperative Agents Safeguarding LLMs Against Evolving Multi-Round AttacksLLM agents and tool use → Agent threats, guardrails and governance
Humans Integrate, Agents Fix: How Agent-Authored Pull Requests Are Referenced in PracticeLLM agents and tool use → Software engineering coding agents
TORA: Topological Representation Alignment for 3D Shape AssemblyDiffusion models → 3D, point cloud and scene generation
Emergent Inference-Time Semantic Contamination via In-Context PrimingLLM safety and security → Safety alignment, refusal and guardrails
Geometric Limits of Knowledge Distillation: A Minimum-Width Theorem via Superposition TheoryKnowledge distillation → General knowledge distillation methods
Causality Laundering: Denial-Feedback Leakage in Tool-Calling LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance
Jellyfish: Zero-Shot Federated Unlearning Scheme with Knowledge DisentanglementLLM safety and security → Federated unlearning and poisoning
1.x-Distill: Breaking the Diversity, Quality, and Efficiency Barrier in Distribution Matching DistillationDiffusion models → Diffusion foundations and sampling
GeoBrowse: A Geolocation Benchmark for Agentic Tool Use with Expert-Annotated Reasoning TracesLLM reasoning → Spatial and embodied reasoning +1
A Systematic Study of Cross-Modal Typographic Attacks on Audio-Visual ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Can LLMs Learn to Reason Robustly under Noisy Supervision?LLM reasoning → Reinforcement learning for reasoning
Gram-Anchored Prompt Learning for Vision-Language Models via Second-Order StatisticsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Predict, Don't React: Value-Based Safety Forecasting for LLM StreamingLLM safety and security → Safety alignment, refusal and guardrails
VLA-Forget: Vision-Language-Action Unlearning for Embodied Foundation ModelsVision-language-action models → Generalist VLA models and benchmarks +1
SafeCtrl: Region-Aware Safety Control for Text-to-Image Diffusion via Detect-Then-SuppressDiffusion models → Robustness, privacy and adversarial aspects +1
What Do We Need for an Agentic Society?LLM agents and tool use → General LLM agent frameworks and benchmarks
CODE-GEN: A Human-in-the-Loop RAG-Based Agentic AI System for Multiple-Choice Question GenerationRetrieval-augmented generation → Agentic RAG, agents and memory +1
Interpreting Video Representations with Spatio-Temporal Sparse AutoencodersMechanistic interpretability → Interpreting vision and multimodal models
Automating Cloud Security and Forensics Through a Secure-by-Design Generative AI FrameworkLLM safety and security → Prompt injection and agent security
Optimal Adaptive Market Making: A Theoretical Framework for High-Yield Liquidity Provision in Perpetual Futures Markets
A Reproducible Semantic Benchmark for Multivendor DSM-to-CLI Translation
Advanced Machine Learning and Deep Learning Techniques for Enhanced Cattle Identification and Detection: A Comprehensive Review
AnonShield: Scalable On-Premise Pseudonymization for CSIRT Vulnerability Data
Towards Next-Generation Healthcare: A Survey of Medical Embodied AI for Perception, Decision-Making, and Action
NeuroSymbolic AI for Legal AI-TRISM: Trustworthy, Reliable, Interpretable, Safe Models
MedicalBench: Evaluating Large Language Models Toward Improved Medical Concept Extraction
Data Scaling as Progressive Coverage of a Predictive Contribution Spectrum
Interoceptive Divergence in Aesthetic Evaluation and Implications for Human-AI Alignment
2026 Roadmap on Artificial Intelligence and Machine Learning for Smart Manufacturing
Simulating Validity: Modal Decoupling in MLLM Generated Feedback on Science Drawings
Non-Destructive Prediction of Fruit Ripeness and Firmness Using Hyperspectral Imaging and Lightweight Machine Learning Models
Conformal PM2.5 Mapping Under Spatial Covariate Shift: Satellite-Reanalysis Fusion for Africa's Green Industrial Transition
Towards Verified and Targeted Explanations through Formal Methods
A Modular Zero-Shot Pipeline for Accident Detection, Localization, and Classification in Traffic Surveillance Video
Learning Stable Predictors from Weak Supervision under Distribution Shift
Closed-Loop Autonomous Software Development via Jira-Integrated Backlog Orchestration: A Case Study in Deterministic Control and Safety-Constrained Automation
PRIME: Prototype-Driven Multimodal Pretraining for Cancer Prognosis with Missing Modalities
El Nino Prediction Based on Weather Forecast and Geographical Time-series Data
Evaluation of Embedding-Based and Generative Methods for LLM-Driven Document Classification: Opportunities and Challenges
Learning-Based Multi-Criteria Decision Making Model for Sawmill Location Problems
The Hiremath Early Detection (HED) Score: A Measure-Theoretic Evaluation Standard for Temporal Intelligence
Prune-Quantize-Distill: An Ordered Pipeline for Efficient Neural Network Compression
Effects of Generative AI Errors on User Reliance Across Task Difficulty
How Long short-term memory artificial neural network, synthetic data, and fine-tuning improve the classification of raw EEG data
Convolutional Neural Network and Adversarial Autoencoder in EEG images classification
Out-of-Air Computation: Enabling Structured Function Extraction from Wireless Superposition
HighFM: Towards a Foundation Model for Learning Representations from High-Frequency Earth Observation Data
CavMerge: Merging K-means Based on Local Log-Concavity
A Persistent Homology Design Space for 3D Point Cloud Deep Learning
PanLUNA: An Efficient and Robust Query-Unified Multimodal Model for Edge Biosignal Intelligence
Adaptive Cost-Efficient Evaluation for Reliable Patent Claim Generation
DAGAF: A directed acyclic generative adversarial framework for joint structure learning and tabular data synthesis
Poisoned Identifiers Survive LLM Deobfuscation: A Case Study on Claude Opus 4.6
Entropy, Disagreement, and the Limits of Foundation Models in Genomics

All 196