{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/chatvla-unified-multimodal-understanding-and","title":"ChatVLA: Unified Multimodal Understanding and Robot Control with Vision-Language-Action Model","arxiv_id":"2502.14420","date":"2025-02-20","proceeding":null,"authors":["Zhongyi Zhou","Yichen Zhu","Minjie Zhu","Junjie Wen","Ning Liu","Zhiyuan Xu","Weibin Meng","Ran Cheng","Yaxin Peng","Chaomin Shen","Feifei Feng"],"abstract":"Humans possess a unified cognitive ability to perceive, comprehend, and interact with the physical world. Why can't large language models replicate this holistic understanding? Through a systematic analysis of existing training paradigms in vision-language-action models (VLA), we identify two key challenges: spurious forgetting, where robot training overwrites crucial visual-text alignments, and task interference, where competing control and understanding tasks degrade performance when trained jointly. To overcome these limitations, we propose ChatVLA, a novel framework featuring Phased Alignment Training, which incrementally integrates multimodal data after initial control mastery, and a Mixture-of-Experts architecture to minimize task interference. ChatVLA demonstrates competitive performance on visual question-answering datasets and significantly surpasses state-of-the-art vision-language-action (VLA) methods on multimodal understanding benchmarks. Notably, it achieves a six times higher performance on MMMU and scores 47.2% on MMStar with a more parameter-efficient design than ECoT. Furthermore, ChatVLA demonstrates superior performance on 25 real-world robot manipulation tasks compared to existing VLA methods like OpenVLA. Our findings highlight the potential of our unified framework for achieving both robust multimodal understanding and effective robot control.","url_abs":"https://arxiv.org/abs/2502.14420v2","url_pdf":"https://arxiv.org/pdf/2502.14420v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"chatvla-unified-multimodal-understanding-and","repo_url":"https://github.com/tutujingyugang1/ChatVLA_public","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"mixture-of-experts","task_name":"Mixture-of-Experts"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"robot-manipulation","task_name":"Robot Manipulation"},{"task_slug":"vision-language-action","task_name":"Vision-Language-Action"},{"task_slug":"visual-question-answering-1","task_name":"Visual Question Answering"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2502.14420","atlas_url":"https://app.syntology.ai/?focus=2502.14420","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}