Papers

New on arXiv

Sun 8 Jun 2025

207 papers

ZettaLith: An Architectural Exploration of Extreme-Scale AI Inference AccelerationEfficient LLM inference and serving → Serving systems and request scheduling
Adaptive Batch-Wise Sample Scheduling for Direct Preference OptimizationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Training-free LLM Verification via Recycling Few-shot ExamplesLLM reasoning → Test-time compute and efficient thinking
Improving Prediction Certainty Estimation for Reliable Early Exiting via Null Space ProjectionEfficient LLM inference and serving → Early exit and edge inference
Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human ValuesLLM agents and tool use → Agent threats, guardrails and governance
History-Aware Cross-Attention Reinforcement: Self-Supervised Multi Turn and Chain-of-Thought Fine-Tuning with vLLMLLM reasoning → Reasoning for agents, planning and code
From Tool Calling to Symbolic Thinking: LLMs in a Persistent Lisp Metaprogramming LoopLLM reasoning → Reasoning for agents, planning and code +1
Enhancing the Safety of Medical Vision-Language Models by Synthetic DemonstrationsMultimodal LLMs / vision-language models → Medical vision-language models
UAVs Meet Agentic AI: A Multidomain Survey of Autonomous Aerial Intelligence and Agentic UAVsLLM agents and tool use → Multimodal and embodied agents
Reward Model Interpretability via Optimal and Pessimal TokensLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
SCGAgent: Recreating the Benefits of Reasoning Models for Secure Code Generation with Agentic WorkflowsLLM agents and tool use → Agent threats, guardrails and governance
Paged Attention Meets FlexAttention: Unlocking Long-Context Efficiency in Deployed InferenceEfficient LLM inference and serving → KV cache optimization
ConfRAG: Confidence-Guided Retrieval-Augmenting GenerationRetrieval-augmented generation → Hallucination detection in RAG
Multi-Step Guided Diffusion for Image Restoration on Edge Devices: Toward Lightweight Perception in Embodied AIDiffusion models → Inverse problems and image restoration
From Generation to Generalization: Emergent Few-Shot Learning in Video Diffusion ModelsDiffusion models → Video generation
Overclocking LLM Reasoning: Monitoring and Controlling Thinking Path Lengths in LLMsLLM reasoning → Test-time compute and efficient thinking
Multi-Step Visual Reasoning with Visual Tokens Scaling and VerificationLLM reasoning → Multimodal reasoning +2
Learn as Individuals, Evolve as a Team: Multi-agent LLMs Adaptation in Embodied EnvironmentsLLM reasoning → Reasoning for agents, planning and code +1
Hallucination at a Glance: Controlled Visual Edits and Fine-Grained Multimodal LearningMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Reflex First, Reflect Later: Latency-Aware Embodied LLM Agents for Dynamic ResponseLLM agents and tool use → Multimodal and embodied agents
Perception-R1: Advancing Multimodal Reasoning Capabilities of MLLMs via Visual Perception RewardLLM reasoning → Multimodal reasoning +1
BIMgent: Towards Autonomous Building Modeling via Computer-use AgentsLLM agents and tool use → Multimodal and embodied agents
Backdoor Attack on Vision Language Models with Stealthy Semantic ManipulationMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
TV-LiVE: Training-Free, Text-Guided Video Editing via Layer Informed Vitality ExploitationDiffusion models → Video generation
Reasoning Multimodal Large Language Model: Data Contamination and Dynamic EvaluationLLM reasoning → Multimodal reasoning +1
Audio synthesizer inversion in symmetric parameter spaces with approximately equivariant flow matchingDiffusion models → Audio, speech and music
SAP-Bench: Benchmarking Multimodal Large Language Models in Surgical Action PlanningMultimodal LLMs / vision-language models → Medical vision-language models
Mitigating Behavioral Hallucination in Multimodal Large Language Models for Sequential ImagesMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Flattery in Motion: Benchmarking and Analyzing Sycophancy in Video-LLMsLLM reasoning → Multimodal reasoning
Frame Guidance: Training-Free Guidance for Frame-Level Control in Video Diffusion ModelsDiffusion models → Video generation
RULE: Reinforcement UnLEarning Achieves Forget-Retain Pareto OptimalityLLM safety and security → Machine unlearning
AMoPO: Adaptive Multi-objective Preference Optimization without Reward Models and Reference ModelsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
GeometryZero: Advancing Geometry Solving via Group Contrastive Policy OptimizationLLM reasoning → Spatial and embodied reasoning
Mind the Web: The Security of Web Use AgentsLLM agents and tool use → Agent threats, guardrails and governance
Prompting Science Report 2: The Decreasing Value of Chain of Thought in PromptingLLM reasoning → Prompting and chain-of-thought reasoning
Learning Compact Vision Tokens for Efficient Large Multimodal ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Human-assisted Robotic Policy Refinement via Action Preference OptimizationVision-language-action models → Reinforcement learning and post-training +1
Quality-Diversity Red-Teaming: Automated Generation of High-Quality and Diverse Attackers for Large Language ModelsLLM safety and security → Prompt injection and agent security
BRIGHT+: Upgrading the BRIGHT Benchmark with MARCUS, a Multi-Agent RAG Clean-Up SuiteRetrieval-augmented generation → Agentic RAG, agents and memory +1
Theorem-of-Thought: A Multi-Agent Framework for Abductive, Deductive, and Inductive Reasoning in Language ModelsLLM reasoning → Reasoning benchmarks and evaluation +1
How Far Are We from Optimal Reasoning Efficiency?LLM reasoning → Test-time compute and efficient thinking
Filling the Missings: Spatiotemporal Data Imputation by Conditional DiffusionDiffusion models → Flow matching, time series and science
SceneLCM: End-to-End Layout-Guided Interactive Indoor Scene Generation with Latent Consistency ModelDiffusion models → 3D, point cloud and scene generation
State Entropy Regularization for Robust Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
Inverse Design of Metamaterials with Manufacturing-Guiding Spectrum-to-Structure Conditional Diffusion ModelDiffusion models → Flow matching, time series and science
Dual-Priv Pruning : Efficient Differential Private Fine-Tuning in Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Reasoning Paths as Signals: Augmenting Multi-hop Fact Verification through Structural Reasoning ProgressionLLM reasoning → Reasoning benchmarks and evaluation
Com2: A Causal-Guided Benchmark for Exploring Complex Commonsense Reasoning in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
A Layered Self-Supervised Knowledge Distillation Framework for Efficient Multimodal Learning on the EdgeKnowledge distillation → General knowledge distillation methods
Policy Gradient with Tree Search: Avoiding Local Optimas through LookaheadDeep reinforcement learning methods → RL theory and sample complexity
QForce-RL: Quantized FPGA-Optimized Reinforcement Learning Compute EngineDeep reinforcement learning methods → RL training of language and agent models
Interpretable and Reliable Detection of AI-Generated Images via Grounded Reasoning in MLLMsLLM reasoning → Multimodal reasoning
Lingshu: A Generalist Foundation Model for Unified Multimodal Medical Understanding and ReasoningLLM reasoning → Clinical and biomedical applications +1
Reasoning with RAGged events: RAG-Enhanced Event Knowledge Base Construction and reasoning with proof-assistantsLLM reasoning → Knowledge- and tool-augmented reasoning
Efficient Q-Learning and Actor-Critic Methods for Robust Average-Reward Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
KG2QA: Knowledge Graph-enhanced Retrieval-augmented Generation for Communication Standards Question AnsweringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
AnnoDPO: Protein Functional Annotation Learning with Direct Preference OptimizationLLM alignment / RLHF / preference optimisation → Preference optimization for molecular design
A Culturally-diverse Multilingual Multimodal Video Benchmark & ModelMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
HauntAttack: When Attack Follows Reasoning as a ShadowLLM reasoning → Reasoning for agents, planning and code
AlphaSteer: Learning Refusal Steering with Principled Null-Space ConstraintLLM safety and security → Safety alignment, refusal and guardrails +1
MAGNET: A Multi-agent Framework for Finding Audio-Visual Needles by Reasoning over Multi-Video HaystacksMultimodal LLMs / vision-language models → Video understanding with MLLMs
CARoL: Context-aware Adaptation for Robot LearningDeep reinforcement learning methods → Deep RL algorithms and representations
What makes Reasoning Models Different? Follow the Reasoning Leader for Efficient DecodingLLM reasoning → Test-time compute and efficient thinking
Certified Unlearning for Neural NetworksLLM safety and security → Machine unlearning
Chain of Methodologies: Scaling Test Time Computation without TrainingLLM reasoning → Prompting and chain-of-thought reasoning
Deep RL Needs Deep Behavior Analysis: Exploring Implicit Planning by Model-Free Agents in Open-Ended EnvironmentsDeep reinforcement learning methods → Deep RL algorithms and representations
Atomic Reasoning for Scientific Table Claim VerificationLLM reasoning → Knowledge- and tool-augmented reasoning
Chain-of-Code Collapse: Reasoning Failures in LLMs via Adversarial Prompting in Code GenerationLLM reasoning → Prompting and chain-of-thought reasoning +1
Guiding Cross-Modal Representations with MLLM Priors via Preference AlignmentLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Offline RL by Reward-Weighted Fine-Tuning for Conversation OptimizationLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Deontically Constrained Policy Improvement in Reinforcement Learning AgentsDeep reinforcement learning methods → RL theory and sample complexity
Simulating Society Requires Simulating ThoughtLLM reasoning → Reasoning for agents, planning and code +1
BIS Reasoning 1.0: The First Large-Scale Japanese Benchmark for Belief-Inconsistent Syllogistic ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Safety-Aware Reinforcement Learning for Control via Risk-Sensitive Action-Value Iteration and Quantile RegressionDeep reinforcement learning methods → Safe and constrained RL
PGR-DRC: Pre-Global Routing DRC Violation Prediction Using Unsupervised Learning
UltraSketchLLM: Sub-1-Bit LLM Compression via Sketch and Hardware-Friendly Operators
Near-Optimal Online Deployment and Routing for Streaming LLMs
MS-DFTVNet:A Long-Term Time Series Prediction Method Based on Multi-Scale Deformable Convolution
Towards Interpretable Adversarial Examples via Sparse Adversarial Attack
Efficient Quantification of Multimodal Interaction at Sample Level
Next-Generation Conflict Forecasting: Unleashing Predictive Patterns through Spatiotemporal Learning
A Hybrid ConvNeXt-EfficientNet AI Solution for Precise Falcon Disease Detection
Predicting Anthropometric Body Composition Variables Using 3D Optical Imaging and Machine Learning
Mirror Descent Using the Tempesta Generalized Multi-parametric Logarithms
Representing Time-Continuous Behavior of Cyber-Physical Systems in Knowledge Graphs
Latency Optimization for Wireless Federated Learning in Multihop Networks
Breaking the Reviewer: Assessing the Vulnerability of Large Language Models in Automated Peer Review Under Textual Adversarial Attacks
Dagstuhl Perspectives Workshop 24352 -- Conversational Agents: A Framework for Evaluation (CAFE): Manifesto
Evaluating and Improving Robustness in Large Language Models: A Survey and Future Directions
AssertBench: A Benchmark for Evaluating Self-Assertion in Large Language Models
Enhancing Large Language Models for Mobility Analytics with Semantic Location Tokenization
Devanagari Digit Recognition using Quantum Machine Learning
BG-HOP: A Bimanual Generative Hand-Object Prior
ReStNet: A Reusable & Stitchable Network for Dynamic Adaptation on IoT Devices
Physics-Informed Teleconnection-Aware Transformer for Global Subseasonal-to-Seasonal Forecasting
Toward Reliable AR-Guided Surgical Navigation: Interactive Deformation Modeling with Data-Driven Biomechanics and Prompts
Evaluation of Machine Learning Models in Student Academic Performance Prediction
Mobility-Aware Asynchronous Federated Learning with Dynamic Sparsification
CASE: Contrastive Activation for Class-Sensitive Explanations
DEF: Diffusion-augmented Ensemble Forecasting

All 207