![]() ![]() Kim, Jong-Seon Kim, Yei Hwan Jung, Tae-il Kim, Cassian Yee, John A. Gayoung Park, Hyun-Joong Chung, Kwanghee Kim, Seon Ah Lim, Jiyoung Kim, Yun-Soung Kim, Yuhao Liu, Woon-Hong Yeo, Rak-Hwan Kim, Stanley S. ![]() Huang X, Cheng H, Chen K, Zhang Y, Zhang Y, Liu Y, Zhu C, Ouyang SC, Kong GW, Yu C, Huang Y, Rogers J, “ Epidermal Impedance Sensing Sheets for Precision Hydration Assessment and Spatial Mapping”, IEEE Trans Biomed Eng.Liu, “ Epidermal Differential Impedance Sensor for Conformal Skin Hydration Monitoring”, Biointerphases (2012) 7:52. Birgeneau, “ Soft X-ray absorption spectroscopy investigations of 1111 and 122 iron pnictides”, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, Volume 649, Issue 1, 1 September 2011, Pages 197-199 Byron Freelon, Yuhao Liu, YS Liu, CR Rotundu, SD Wilson, JH Guo, Jenglung Chen, W Yang, CL Chang, PA Glans, PM Shirage, A Iyo, RJ Birgeneau, “ Electronic Structure of PrFeAsO1− δ: An Investigation Using X-ray Absorption and Emission Spectroscopy“, Journal of Physics: Conference Series, 1 January 2011. ![]() We further validated our method on a real robot arm where complex manipulation skills such as non-prehensile pushing emerge through our interactive system.Journal Publications: (Total Citations: 273, h-index: 9 Google Scholar 11/2015) We demonstrate that our proposed method reliably tackles 90% of the designed tasks, while a baseline using primitive skills as the interface with Code-as-policies achieves 50% of the tasks. To systematically evaluate the performance of our proposed method, we designed a total of 17 tasks for a simulated quadruped robot and a dexterous manipulator robot. Meanwhile, combining this with a real-time optimizer, MuJoCo MPC, empowers an interactive behavior creation experience where users can immediately observe the results and provide feedback to the system. Using reward as the intermediate interface generated by LLMs, we can effectively bridge the gap between high-level language instructions or corrections to low-level robot actions. In this work, we introduce a new paradigm that harnesses this realization by utilizing LLMs to define reward parameters that can be optimized and accomplish variety of robotic tasks. On the other hand, reward functions are shown to be flexible representations that can be optimized for control policies to achieve diverse tasks, while their semantic richness makes them suitable to be specified by LLMs. However, since low-level robot actions are hardware-dependent and underrepresented in LLM training corpora, existing efforts in applying LLMs to robotics have largely treated LLMs as semantic planners or relied on human-engineered control primitives to interface with the robot. Robotics researchers have also explored using LLMs to advance the capabilities of robotic control. Download a PDF of the paper titled Language to Rewards for Robotic Skill Synthesis, by Wenhao Yu and 19 other authors Download PDF Abstract:Large language models (LLMs) have demonstrated exciting progress in acquiring diverse new capabilities through in-context learning, ranging from logical reasoning to code-writing. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |