Conditionally Combining Robot Skills using Large Language Models

被引:0
|
作者
Zentner, K. R. [1 ,3 ]
Julian, Ryan [2 ]
Ichter, Brian [2 ]
Sukhatme, Gaurav S. [1 ]
机构
[1] Univ Southern Calif, Los Angeles, CA 90007 USA
[2] Google DeepMind, London, England
[3] Google Brain, Mountain View, CA USA
来源
2024 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2024) | 2024年
关键词
D O I
10.1109/ICRA57147.2024.10611275
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This paper combines two contributions. First, we introduce an extension of the Meta-World benchmark, which we call "Language-World," which allows a large language model to operate in a simulated robotic environment using semi-structured natural language queries and scripted skills described using natural language. By using the same set of tasks as Meta-World, Language-World results can be easily compared to Meta-World results, allowing for a point of comparison between recent methods using Large Language Models (LLMs) and those using Deep Reinforcement Learning. Second, we introduce a method we call Plan Conditioned Behavioral Cloning (PCBC), that allows finetuning the behavior of high-level plans using end-to-end demonstrations. Using Language-World, we show that PCBC is able to achieve strong performance in a variety of few-shot regimes, often achieving task generalization with as little as a single demonstration. We have made Language-World available as open-source software at https://github.com/krzentner/language-world/.
引用
收藏
页码:14046 / 14053
页数:8
相关论文
共 50 条
  • [1] Generative Expressive Robot Behaviors using Large Language Models
    Mahadevan, Karthik
    Chien, Jonathan
    Brown, Noah
    Xu, Zhuo
    Parada, Carolina
    Xia, Fei
    Zeng, Andy
    Takayama, Leila
    Sadigh, Dorsa
    PROCEEDINGS OF THE 2024 ACM/IEEE INTERNATIONAL CONFERENCE ON HUMAN-ROBOT INTERACTION, HRI 2024, 2024, : 482 - 491
  • [2] Generation of Robot Manipulation Plans Using Generative Large Language Models
    Toberg, Jan-Philipp
    Cimiano, Philipp
    2023 SEVENTH IEEE INTERNATIONAL CONFERENCE ON ROBOTIC COMPUTING, IRC 2023, 2023, : 190 - 197
  • [3] A framework for neurosymbolic robot action planning using large language models
    Capitanelli, Alessio
    Mastrogiovanni, Fulvio
    FRONTIERS IN NEUROROBOTICS, 2024, 18
  • [4] A Survey of Robot Intelligence with Large Language Models
    Jeong, Hyeongyo
    Lee, Haechan
    Kim, Changwon
    Shin, Sungtae
    APPLIED SCIENCES-BASEL, 2024, 14 (19):
  • [5] PROGPROMPT: Generating Situated Robot Task Plans using Large Language Models
    Singh, Ishika
    Blukis, Valts
    Mousavian, Arsalan
    Goyal, Ankit
    Xu, Danfei
    Tremblay, Jonathan
    Fox, Dieter
    Thomason, Jesse
    Garg, Animesh
    2023 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2023), 2023, : 11523 - 11530
  • [6] TidyBot: Personalized Robot Assistance with Large Language Models
    Wu, Jimmy
    Antonova, Rika
    Kan, Adam
    Lepert, Marion
    Zeng, Andy
    Song, Shuran
    Bohg, Jeannette
    Rusinkiewicz, Szymon
    Funkhouser, Thomas
    2023 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS, IROS, 2023, : 3546 - 3553
  • [7] TidyBot: personalized robot assistance with large language models
    Wu, Jimmy
    Antonova, Rika
    Kan, Adam
    Lepert, Marion
    Zeng, Andy
    Song, Shuran
    Bohg, Jeannette
    Rusinkiewicz, Szymon
    Funkhouser, Thomas
    AUTONOMOUS ROBOTS, 2023, 47 (08) : 1087 - 1102
  • [8] TidyBot: personalized robot assistance with large language models
    Jimmy Wu
    Rika Antonova
    Adam Kan
    Marion Lepert
    Andy Zeng
    Shuran Song
    Jeannette Bohg
    Szymon Rusinkiewicz
    Thomas Funkhouser
    Autonomous Robots, 2023, 47 : 1087 - 1102
  • [9] Evaluation of radiology residents' reporting skills using large language models: an observational study
    Atsukawa, Natsuko
    Tatekawa, Hiroyuki
    Oura, Tatsushi
    Matsushita, Shu
    Horiuchi, Daisuke
    Takita, Hirotaka
    Mitsuyama, Yasuhito
    Omori, Ayako
    Shimono, Taro
    Miki, Yukio
    Ueda, Daiju
    JAPANESE JOURNAL OF RADIOLOGY, 2025,
  • [10] Using Large Language Models To Diagnose Math Problem-solving Skills At Scale
    Jin, Hyoungwook
    Kim, Yoonsu
    Park, Yeon Su
    Tilekbay, Bekzat
    Son, Jinho
    Kim, Juho
    PROCEEDINGS OF THE ELEVENTH ACM CONFERENCE ON LEARNING@SCALE, L@S 2024, 2024, : 471 - 475