Discovering Generalizable Skills via Automated Generation of Diverse Tasks

被引:0
|
作者
Fang, Kuan [1 ]
Zhu, Yuke [2 ,3 ]
Savarese, Silvio [1 ]
Li Fei-Fei [1 ]
机构
[1] Stanford Univ, Stanford, CA 94305 USA
[2] UT Austin, Austin, TX USA
[3] Nvidia, Santa Clara, CA USA
关键词
REINFORCEMENT; OBJECTS;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The learning efficiency and generalization ability of an intelligent agent can be greatly improved by utilizing a useful set of skills. However, the design of robot skills can often be intractable in real-world applications due to the prohibitive amount of effort and expertise that it requires. In this work, we introduce Skill Learning In Diversified Environments (SLIDE), a method to discover generalizable skills via automated generation of a diverse set of tasks. As opposed to prior work on unsupervised discovery of skills which incentivizes the skills to produce different outcomes in the same environment, our method pairs each skill with a unique task produced by a trainable task generator. To encourage generalizable skills to emerge, our method trains each skill to specialize in the paired task and maximizes the diversity of the generated tasks. A task discriminator defined on the robot behaviors in the generated tasks is jointly trained to estimate the evidence lower bound of the diversity objective. The learned skills can then be composed in a hierarchical reinforcement learning algorithm to solve unseen target tasks. We demonstrate that the proposed method can effectively learn a variety of robot skills in two tabletop manipulation domains. Our results suggest that the learned skills can effectively improve the robot's performance in various unseen target tasks compared to existing reinforcement learning and skill learning methods.
引用
收藏
页数:12
相关论文
共 50 条
  • [41] ShellFusion: Answer Generation for Shell Program ming Tasks via Knowledge Fusion
    Zhang, Neng
    Liu, Chao
    Xia, Xin
    Treude, Christoph
    Zou, Ying
    Lo, David
    Zheng, Zibin
    2022 ACM/IEEE 44TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE 2022), 2022, : 1970 - 1981
  • [42] 3DRobot: automated generation of diverse and well-packed protein structure decoys
    Deng, Haiyou
    Jia, Ya
    Zhang, Yang
    BIOINFORMATICS, 2016, 32 (03) : 378 - 387
  • [43] MixPoet: Diverse Poetry Generation via Learning Controllable Mixed Latent Space
    Yi, Xiaoyuan
    Li, Ruoyu
    Yang, Cheng
    Li, Wenhao
    Sun, Maosong
    THIRTY-FOURTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THE THIRTY-SECOND INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE AND THE TENTH AAAI SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2020, 34 : 9450 - 9457
  • [44] DivGAN: Towards Diverse Paraphrase Generation via Diversified Generative Adversarial Network
    Cao, Yue
    Wan, Xiaojun
    FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, EMNLP 2020, 2020, : 2411 - 2421
  • [45] Generation of optical vortices with diverse topological charge via angular momentum transfer
    Lopez-Quintas, Ignacio
    Holgado, Warein
    Drevinskas, Rokas
    Kazansky, Peter G.
    Sola, Inigo J.
    Alonso, Benjamin
    2021 CONFERENCE ON LASERS AND ELECTRO-OPTICS EUROPE & EUROPEAN QUANTUM ELECTRONICS CONFERENCE (CLEO/EUROPE-EQEC), 2021,
  • [46] BioGraph: unsupervised biomedical knowledge discovery via automated hypothesis generation
    Liekens, Anthony M. L.
    De Knijf, Jeroen
    Daelemans, Walter
    Goethals, Bart
    De Rijk, Peter
    Del-Favero, Jurgen
    GENOME BIOLOGY, 2011, 12 (06):
  • [47] BioGraph: unsupervised biomedical knowledge discovery via automated hypothesis generation
    Anthony ML Liekens
    Jeroen De Knijf
    Walter Daelemans
    Bart Goethals
    Peter De Rijk
    Jurgen Del-Favero
    Genome Biology, 12
  • [48] Talk2Face: A Unified Sequence-based Framework for Diverse Face Generation and Analysis Tasks
    Li, Yudong
    Hou, Xianxu
    Zhao, Zhe
    Shen, Linlin
    Yang, Xuefeng
    Yan, Kimmo
    PROCEEDINGS OF THE 30TH ACM INTERNATIONAL CONFERENCE ON MULTIMEDIA, MM 2022, 2022, : 4594 - 4604
  • [49] Feasibility of a semi-automated approach to grading point of care ultrasound image generation skills
    Chen, Zizui
    Shehata, Mohamed S.
    Gong, Minglun
    Carnahan, Heather
    Dubrowski, Adam
    Smith, Andrew
    2015 INTERNATIONAL CONFERENCE ON IMAGE AND VISION COMPUTING NEW ZEALAND (IVCNZ), 2015,
  • [50] Generalizable Pancreas Segmentation Modeling in CT Imaging via Meta-Learning and Latent-Space Feature Flow Generation
    Li, Jun
    Chen, Tao
    Qian, Xiaohua
    IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS, 2023, 27 (01) : 374 - 385