Dongyoung Kim

Verifier-free Test-Time Sampling for Vision-Language-Action Models

SpatialBoost: Enhancing Visual Representation through Language-Guided Reasoning

Contrastive Representation Regularization for Vision-Language-Action Models

RoboAlign: Reinforcement Learning for Action-Aligned Multimodal Large Language Models

Training-free LLM Verification via Recycling Few-shot Examples

Dual-Stream Diffusion for World-Model Augmented Vision-Language-Action Model

INFUSER: Injecting Synthetic Failures for Self-Correcting Embodied Agents

Efficient LLM Collaboration via Planning

Spread Preference Annotation: Direct Preference Judgment for Efficient LLM Alignment

Robot-R1: Reinforcement Learning for Enhanced Embodied Reasoning in Robotics

Debiasing Online Preference Learning via Preference Feature Preservation