Yaodong Yang

PoliCon: Evaluating LLMs on Achieving Diverse Political Consensus Objectives

On Robustness of Vision-Language-Action Model against Multi-Modal Perturbations

Align Once, Benefit Multilingually: Enforcing Multilingual Consistency for LLM Safety Alignment

MVR: Multi-view Video Reward Shaping for Reinforcement Learning

Vulnerable Agent Identification in Large-Scale Multi-Agent Reinforcement Learning

Adversarial Policy Transfer in Mixed Cooperative-Competitive Games

Dexterous Non-Prehensile Manipulation for Ungraspable Objects via Extrinsic Dexterity

SIV-Bench: A Video Benchmark for Social Interaction Understanding and Reasoning

Debate with Image: Detecting Deceptive Behaviors in Multimodal Large Language Models

Mesa and Mask: A Benchmark for Detecting and Classifying Deceptive Behaviors in LLMs

Fault Tolerant Multi-Agent Learning with Adversarial Budget Constraints

Safe Reinforcement Learning with ADRC Lagrangian Method

Communication-Efficient Desire Alignment for Embodied Agent–Human Adaptation

SafeEditor: Unified MLLM for Efficient Post-hoc T2I Safety Editing

Mitigating Deceptive Alignment via Self-Monitoring

SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning

Generative RLHF-V: Learning Principles from Multi-modal Human Preference

DexFlyWheel: A Scalable and Self-improving Data Generation Framework for Dexterous Manipulation

Social World Model-Augmented Mechanism Design Policy Learning

Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback

Risk-aware Direct Preference Optimization under Nested Risk Measure

In-Context Editing: Learning Knowledge from Self-Induced Distributions

Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization

Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Model Alignment

STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization

Empirical Study on Robustness and Resilience in Cooperative Multi-Agent Reinforcement Learning

Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMs

SAE-V: Interpreting Multimodal Models for Enhanced Alignment

Falcon: Fast Visuomotor Policies via Partial Denoising

Emerging Safety Attack and Defense in Federated Instruction Tuning of Large Language Models

Risk-aware Direct Preference Optimization under Nested Risk Measure

Random Feature Models with Learnable Activation Functions

Iterative Training of Language Models with Opponent Modeling for Red Teaming Data Generation

Computing Ex Ante Equilibrium in Heterogeneous Zero-Sum Team Games

Mixed Hierarchical Oracle and Multi-Agent Benchmark in Two-player Zero-sum Games