11 papers
WAM-OPD: Joint Video-Action Supervision for World Action Model Post-Training with On-Policy Distillation
Liuhaichen Yang, Zhengyang Zhong, Zhuang Jiang +13
World Action Models (WAMs) generate both future video and robot actions, offering two connected outputs for post-training supervision. How can a pretrained WAM learn from a stronge…
Guided Action Flow: Value-Guided Sampling for Frozen Vision-Language-Action Policies
Liuhaichen Yang, Zhuang Jiang, Chenchao Sheng +8
Reinforcement learning can improve vision-language-action (VLA) policies beyond supervised fine-tuning, although this typically involves further updates to the policy parameters. F…
SignVLA: Real-Time Sign Language-Guided Robotic Manipulation via Attention LSTM and Vision-Language-Action Models
Ningwei Bai, Xinyu Tan, Harry Gardner +6
Vision-Language-Action (VLA) models enable robots to execute manipulation tasks from natural-language instructions grounded in visual observations. However, existing VLA interfaces…
SignVLA: A Gloss-Free Vision-Language-Action Framework for Real-Time Sign Language-Guided Robotic Manipulation
Xinyu Tan, Ningwei Bai, Harry Gardener +6
We present, to our knowledge, the first sign language-driven Vision-Language-Action (VLA) framework for intuitive and inclusive human-robot interaction. Unlike conventional approac…
Reinforced Attention Learning
Bangzheng Li, Jianmo Ni, Chen Qu +5
Post-training with Reinforcement Learning (RL) has substantially improved reasoning in Large Language Models (LLMs) via test-time scaling. However, extending this paradigm to Multi…
Task Vectors in In-Context Learning: Emergence, Formation, and Benefit
Liu Yang, Ziqian Lin, Kangwook Lee +2
In-context learning is a remarkable capability of transformers, referring to their ability to adapt to specific tasks based on a short history or context. Previous research has fou…
Everything Everywhere All at Once: LLMs can In-Context Learn Multiple Tasks in Superposition
Zheyang Xiong, Ziyang Cai, John Cooper +11
Large Language Models (LLMs) have demonstrated remarkable in-context learning (ICL) capabilities. In this study, we explore a surprising phenomenon related to ICL: LLMs can perform…
RS-DPO: A Hybrid Rejection Sampling and Direct Preference Optimization Method for Alignment of Large Language Models
Saeed Khaki, JinJin Li, Lan Ma +2
Reinforcement learning from human feedback (RLHF) has been extensively employed to align large language models with user intent. However, proximal policy optimization (PPO) based R…
How Well Can Transformers Emulate In-context Newton's Method?
Angeliki Giannou, Liu Yang, Tianhao Wang +2
Transformer-based models have demonstrated remarkable in-context learning capabilities, prompting extensive research into its underlying mechanisms. Recent studies have suggested t…
An Empirical Study on the Power of Future Prediction in Partially Observable Environments
Jeongyeol Kwon, Liu Yang, Robert Nowak +1
Learning good representations of historical contexts is one of the core challenges of reinforcement learning (RL) in partially observable environments. While self-predictive auxili…
Looped Transformers are Better at Learning Learning Algorithms
Liu Yang, Kangwook Lee, Robert Nowak +1
Transformers have demonstrated effectiveness in in-context solving data-fitting problems from various (latent) models, as reported by Garg et al. However, the absence of an inheren…