On the Efficiency of LoRA Fine-Tuning for Vision-Language-Action Models in Industrial Robotic Manipulation
工业机器人操作中视觉语言动作模型的LoRA微调效率研究
Finn Ferchau, Daniel Pommer, Cristian Axenie
机构
*
Technische Hochschule Nürnberg Georg Simon Ohm(纽伦堡乔治·西蒙·欧姆应用技术大学)
;
Siemens AG(西门子股份公司)
;
Fraunhofer Institute for Integrated Circuits (IIS)(弗劳恩霍夫集成电路研究所)
机构
*
School of Information Science, Japan Advanced Institute of Science and Technology(日本北陆先端科学技术大学院大学信息科学学院)
;
University of Engineering and Technology, Vietnam National University(越南国立大学工程与技术大学)
;
Department of Robotics, Hanyang University(汉阳大学机器人学系)
机构
*
Harbin Institute of Technology, China(哈尔滨工业大学,中国)
;
Harbin Institute of Technology (Weihai) Qingdao Research Institute, China(哈尔滨工业大学(威海)青岛研究院,中国)
;
Iray Technology co., Ltd., Shandong, China(Iray科技有限公司,山东,中国)
;
Tsinghua University, Beijing, China(清华大学,北京,中国)
Commentsv2: accepted at IEEE Access (2026); minor revisions per peer review, added WiLoR occlusion-mitigation experiment, error analysis, EMA ablation, and author photos
机构
*
Sun Yat-sen University(中山大学)
;
Pengcheng Laboratory(鹏城实验室)
;
Nanyang Technological University(南洋理工大学)
;
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences(深圳先进技术研究院,中国科学院)
;
X-Era AI Lab(X-Era AI实验室)
KAN We Flow? Advancing Robotic Manipulation with 3D Flow Matching via KAN & RWKV
KAN We Flow? 通过KAN与RWKV实现3D操作的机器人操控进步
Zhihao Chen, Yiyuan Ge, Ziyang Wang
机构
*
School of Intelligent Engineering and Automation, Beijing University of Posts and Telecommunications (BUPT)(北京邮电大学智能工程与自动化学院)
;
Beijing Hydrogen Intelligence Technology Co. Ltd.(北京氢能智能科技有限公司)
;
School of Electronic and Information Engineering, South China University of Technology(华南理工大学电子与信息学院)