Papers

New on arXiv

Sat 12 Sep 2026

218 papers

SiliconBench: Speed, Memory, and Fidelity for LLM Serving on Unified-Memory DesktopsEfficient LLM inference and serving → Serving systems and request scheduling
Not All Agents Are Equal: Code Quality and Post-Merge Maintenance Across Five Autonomous Coding Agents in the WildLLM agents and tool use → Software engineering coding agents
Data-Free On-Policy Distillation: How Far Can We Go Without External Data?Knowledge distillation → On-policy and self-distillation for LLMs
A Multi-Stage Agentic Framework for Effective Counter-Narrative Generation and RefinementLLM agents and tool use → Social simulation and multi-agent debate
CyFM: Cylindrical Optimal Transport for Few-Step Complex-Valued Flow MatchingDiffusion models → Flow matching, time series and science
When Tools Get in the Way: The Effect of Unnecessary Tool Availability on LLM AnsweringLLM agents and tool use → Tool learning and function calling
Signatures of Steerability in Activation Space of Language ModelsMechanistic interpretability → Activation steering methods
One Size Does Not Fit All: Setting Inference Depth from the Questions a Deployment Actually AsksEfficient LLM inference and serving → Early exit and edge inference
LIMBO: Lifelong Inference-Time Memory and Budget Optimization for LLM AgentsLLM agents and tool use → Agent memory systems
RA-CoA: Training-free Fashion Image Captioning via Retrieval-Augmented Chain-of-AttributesRetrieval-augmented generation → Multimodal RAG
SkillSecurer: Detecting and Patching Prompt-Injection Vulnerabilities in AI Agent SkillsLLM agents and tool use → Agent threats, guardrails and governance
SafePG: Safe and Globally Optimal Reinforcement Learning with Hard ConstraintsDeep reinforcement learning methods → Safe and constrained RL
LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World ModelsWorld models → Latent predictive world models
Multi-Modal Tumor Survival Prediction via Graph-Guided Mixture of ExpertsMixture-of-experts → Medical and biomedical MoE
Fixed-Time Resilient Integral Reinforcement Learning for Input-Constrained Unknown Nonlinear Systems Under FDI Attacks and Disturbances: A Data-Driven Admissible Warm StartDeep reinforcement learning methods → Safe and constrained RL
GraMRAG: Orchestrating Multi-Agent Multi-Step Reasoning via Graph Memory with Reinforcement LearningLLM reasoning → Knowledge- and tool-augmented reasoning +1
AGENTQ: Quantization-Conditioned Backdoor Attacks on LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance +1
Data-Efficient Agentic Graph Domain Adaptation via Reliability-Aware Prototype LearningLLM agents and tool use → Web, search, and research agents
Conditional Quantum Flow Matching for Data-Scarce Physiological Signal AugmentationDiffusion models → Medical imaging and segmentation
Confuse the Model, Control the Flow: Understanding and Mitigating Privacy Leakage from LLM Agents with Information Flow ControlLLM agents and tool use → Agent threats, guardrails and governance
Unlocking the Unsolvable: Teacher-Guided Curriculum for Data-Efficient RLVRLLM reasoning → Reinforcement learning for reasoning
Synthetic Data in Marketing Research: How to Evaluate and When to TrustSynthetic data and self-improvement for LLMs → LLM synthetic data generation
P3Rec: Distilling Prior--Posterior Preference Reasoning for LLM-based RecommendationLLM reasoning → LLM reasoning for recommendation
Diffusion-Based Multiple-Shooting Indirect Optimal Control for Fuel-Optimal Spacecraft Trajectory GenerationDiffusion models → Diffusion and flow policies for robotics
A Low-Latency Interactive System for Real-Time Video Understanding Based on VLMsMultimodal LLMs / vision-language models → Video understanding with MLLMs
What Makes an Efficient VLA? Navigating Action-Head Design, Scaling, and LatencyVision-language-action models → VLA architectures and generalization
Rethinking the Implications of Human Feedback for Preference Learning in Human-Robot CollaborationLLM reasoning → Reinforcement learning for reasoning +1
Thought without systematicity? Evaluating reasoning models on rule induction tasksLLM reasoning → Reasoning benchmarks and evaluation
CRITICS - Critical Science Without Borders: Language Models to Promote Critical Thinking in Science EducationLLM reasoning → Reasoning benchmarks and evaluation
BLInD: Learning Driver Intent as a Distribution over Future Ego TrajectoriesDiffusion models → Diffusion and flow policies for robotics
North Small Translate: Advanced Cost-Effective Translation (Cohere CAT+)Mixture-of-experts → MoE large language models
DiTAR+: Dual Optimization for Robust Autoregressive Diffusion Speech SynthesisDiffusion models → Audio, speech and music
When Malicious Instructions Persist: Persistent Memory Poisoning Attack on Harness-Based AgentsLLM agents and tool use → Agent threats, guardrails and governance
PriMobiBench: Characterizing Visual Privacy Leakage in VLM-Driven Mobile GUI AgentsLLM agents and tool use → Agent threats, guardrails and governance
The Filter Metric is Safety-Critical: Phantom Advantages in Group-Relative RL under Shaped RewardsLLM reasoning → Reinforcement learning for reasoning
ClinAgent: A ReAct-Based Agent for Conversational Access to Clinical Trial InformationLLM agents and tool use → Medical and scientific domain agents
ShopEase: A Generative AI-Based Multi-Agent Framework for Intelligent Enterprise Customer Support Using Hybrid Retrieval-Augmented GenerationRetrieval-augmented generation → Agentic RAG, agents and memory
ReWeight: Leveraging Human Data for VLA Post-Training via Demonstration Retrieval and Sample WeightingVision-language-action models → Human-video and latent-action pretraining
UniCAR-RL: Seeing Better before Thinking Deeper in Visual MathematicsLLM reasoning → Multimodal reasoning +1
LePlanner: An Iterative Amortized Controller For World ModelsWorld models → Latent predictive world models
ViperQ: Order Flow Pattern Recognition via Auction Market Theory for Reinforcement Learning TradingDeep reinforcement learning methods → Adversarial robustness and RL in finance
GeomVLA: Unifying Scene, Motion, and Action in 3DVision-language-action models → 3D and spatial-aware VLA
DARE: Dialectical Agentic Reasoning for Structured Knowledge Fact CheckingLLM reasoning → Reasoning benchmarks and evaluation +1
Understanding the Limits of Agentic ICD CodingLLM agents and tool use → Medical and scientific domain agents
LLM-Enhanced Multi-Agent Reinforcement Learning for Unified Electric Vehicles-Charging Station-Grid Optimization in Public Charging SystemsLLM agents and tool use → Financial trading and forecasting agents
StepPrune: Adaptive Sequential Visual Token Selection across Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Force-Aware Reinforcement Learning with Hybrid Sensorless Force Estimation for Wheeled-Legged Loco-ManipulationDeep reinforcement learning methods → RL for robot locomotion and manipulation
Does Reasoning Improve Psychological Depth in Large Language Models? It Depends on Who's JudgingLLM reasoning → Reasoning benchmarks and evaluation
Training Specialist Models without Reasoning Trajectories for Domain Expert DistillationLLM reasoning → Prompting and chain-of-thought reasoning +1
Positioning manuscripts in the scientific landscape with agentic AILLM agents and tool use → Web, search, and research agents
How Many Thoughts Can a Vector Hold? The Capacity of Reasoning by SuperpositionLLM reasoning → Test-time compute and efficient thinking
HarnessBandit: Joint Learnability-Transferability Scheduling for Multi-Harness Agentic Reinforcement LearningLLM agents and tool use → Software engineering coding agents
ForeSight: Enhancing Risk Monitoring via Early Safety Signal DistillationLLM safety and security → Safety alignment, refusal and guardrails
Trustworthy Agentic AI: A Comprehensive Cybersecurity and Systems Survey on Threat Landscapes, Defense Architectures, and Open ChallengesLLM agents and tool use → Agent threats, guardrails and governance
JumpStart Your Policy Learning with Lessons from 160,000 Training RunsOffline RL and imitation learning from data → Offline RL algorithms
Oops, Not Now: PEARL, a RAG-Based Support Agent for Gameplay and What Players Want from AI HelpLLM agents and tool use → Reinforcement learning for LLM agents
GROOVE: Geometry-Guided Reduction of Operational-Space Jerk in VLA ExecutionVision-language-action models → Efficient VLA inference and adaptation
Prefix Sharing Is a Sorting ProblemEfficient LLM inference and serving → KV cache optimization
LayerRoute: Adaptive Layer-Skipping with LoRA-Preserved Quality for Efficient LLM InferenceEfficient LLM inference and serving → Early exit and edge inference
An Efficient and Modular Framework for Targeted Harm Mitigation in LLMSLLM safety and security → Safety alignment, refusal and guardrails
Causal discovery identifies pathways linking physical activity to dementia risk in the UK BioBank
Typed Temporal Interaction Features for Simulation-Backed Forecasting of Open-Source Game Release Incidents
Measurement-Error-Aware Causal Distributed-Lag Quantile Modeling of Indoor Air Pollution and Short-Term Lung-Function Deterioration
STAR: Adaptive Spatial-Temporal Normalization for Unified Microservice Incident Management
MaD-RL: Matching Distributions for Calibrating LLMs with Reinforcement Learning
Information Design Against Gaming and Learning Adversaries
A Survey on Fake Review Detection: From Pre-trained Language Models to Large Language Models
LLMersion: A Local-First AI Agent Framework for Low-Cost Home Language Learning toward Educational Equity
An Affordable AI-Integrated Smart Cane for Multimodal Mobility Assistance of Visually Impaired Users
Prior-Free Competitive Ratios for Improving Bandits: Scale, Curvature and Horizon Are Free, but Not Jointly Under Noise
When the Gradient Sees Rank: Provable Necessity, Causal Recruitment, and Composition in Trained Matrix Memories
MorphoStyle: Motion Style Transfer with Morphology Control
Entropy-Punctured Bloom Filters for Memory-Efficient Machine Learning
Bi-Level Routing and Sparse Spatial Attention based Multi-View BEV 3D Object Detection for Autonomous Driving
ZAPS: Zero-Cost Active Proxy Search for Neural Architecture Search
A Machine Learning Framework for Fault Detection, Isolation, and Severity Prediction of Autonomous VTOL Aircraft
Inherited Heads: Audio language models track speakers with their text backbone's attention, and an attention-mass ranking retrieves a different set
Finite-Sample FDR Control for Greedy Aggregation over Networks
Towards Evolving Context Parameterization for Large Language Models
3D Gait-Based Autism Classification Using Attention-Enhanced Deep Learning with Cross-Fold Statistical Stability Analysis
When Faster VLA Deployment Changes Closed-Loop Behavior: Task Success-Latency Analysis of SmolVLA Across PyTorch and ONNX Variants
A New Transformer-Based Approach for Audio-Based Kinship Verification and a New Uncontrolled Mandarin Kinship Speech Dataset
T-SMART: Mechanism-Level Attribution for Tool-Augmented Time-Series Question Answering
Riemannian ascent--descent for nonconvex nonconcave minimax landscapes: convergence to basin saddle points and applications to distributionally robust optimization
Deformable 2D Gaussian Splatting for Efficient 4K Video Compression
Exact Finite Attention Responses From RoPE Derivatives
To do(x) or not to do(x): Medical Image Counterfactuals for Dataset Augmentation
SignMimic: Robust High-Quality Sign Language Motion Generation via Human-Shape-Oblivious Pose Transfer Guidance
Semantic Knowledge Technologies: what the Semantic Web lost sight of, and what it never had
Talking to Me or Someone Else? Rethinking Talk-to-Me Detection in Egocentric Videos
Real-Time Synthesis of Robust Controlled Invariant Sets for Monotone Systems
Parameter-Efficient Fine-Tuning of Foundation Models for Liver Tumor Segmentation in CT
Accelerating HKTex without Mesh Eigensystems: Local Unfolding and Randomized Thermal Features
A Voxel-Spacing-Aware Extension of PyRadiomics for Anisotropic Texture Analysis
Neuron Activation-based Computation of Logical Explanations for Deep Neural Networks
Bridging the Synthetic-to-Real Gap for Few-Shot Cryo-ET Classification
NeuroFlex: Lossless Element-Level ANN-SNN Co-Execution for Efficient Sparse Inference
Just add noise: Debiasing tree-based variable importance in mixed data
Adapting Open-Weight MLLMs to Generate Point Prompts for Electron Microscopy Segmentation
Stabilizing Performative Feedback Loops with Minimal Model Deployments

All 218