Welcome to the CLVR lab!

We are the Cognitive Learning for Vision and Robotics (CLVR, pronounced as clever) lab at Korea Advanced Institute of Science and Technology (KAIST), in Seoul, led by Joseph J. Lim. Our goal is to develop intelligent systems that make sequential decisions, in the world, through perception, action, and reasoning. The research topics include reinforcement learning, robot learning, world model with representation learning, visual perception, and symbolic manipulation.


Recent News (more)

[Paper] Two papers are accepted to RSS! (2023)
[Talk] Joseph J. Lim gives a talk at CoRL! (2022)
[Paper] Two papers are accepted to CoRL! (2022)
[Paper] Three papers are accepted to ICLR! (2022)

Research Topics (more)



Cognitive Model
(e.g., Domain/Action Generalization)

Robot Learning
(e.g., Complex Long-Horizon Manipulation)

Reinforcement Learning
(e.g., Skill-based Reinforcement Learning)


Logical Reasoning
(e.g., Physics Understanding)

Symbolic Manipulation
(e.g., Neural Program Synthesis)

Computer Vision
(e.g., Learning from Videos)

News

2023
  • [Paper] Two papers are accepted to RSS!
  • [PhD Defense] Karl Pertsch successfully defended his PhD thesis!
2022
  • [Talk] Joseph J. Lim gives a talk at CoRL!
  • [Paper] Two papers are accepted to CoRL!
  • [Paper] Three papers are accepted to ICLR!
  • [PhD Defense] Youngwoon Lee successfully defended his PhD thesis!
  • [PhD Defense] Shao-Hua Sun successfully defended his PhD thesis!
2021
  • [Paper] Two papers are accepted to NeurIPS!
  • [Paper] Three papers are accepted to CoRL!
  • [Paper] Our paper on policy transfer across visual and dynamics gaps is accepted to RSS!
  • [Paper] One paper is accepted to ICML!
  • [Paper] Our IKEA Furniture Assembly Environment is accepted to ICRA!
2020
  • [Award] SPiRL is selected as a plenary talk and won the best paper presentation award at CoRL!
  • [Paper] Two papers are accepted to CoRL!
  • [Paper] Paper on generalization to new actions is accepted to ICML!
  • [Paper] Two papers are accepted to ICLR!
  • [Release] We released IKEA Furniture Assembly Environment for a complex robot manipulation benchmark.
2019
  • [Paper] Our paper on imitation learning from observation is accepted to CoRL!
  • [Paper] Our paper on multimodal meta-learning is accepted to NeurIPS!
  • [Talk] Deepak Pathak from UC Berkeley gave a talk as part of the CLVR Speaker Series.
  • [Paper] Paper on complex skill composition accepted to ICLR 2019!
  • [Talk] Kuan Fang from the Stanford Vision Lab gave a talk as part of the CLVR Speaker Series.
  • [Talk] Tianmin Shu from UCLA gave a talk as part of the CLVR Speaker Series.
  • [Talk] Yuke Zhu from the Stanford Vision Lab gave a talk as part of the CLVR Speaker Series.
  • [Talk] Andrew Owens from UC Berkeley gave a talk as part of the CLVR Speaker Series.
2018
  • [Paper] Paper on novel-view synthesis accepted to ECCV!
  • [Paper] Paper on neural program synthesis accepted to ICML!
  • [Talk] Jiajun Wu from MIT gave a talk as part of the CLVR Speaker Series.
  • [Paper] Paper on object affordance reasoning accepted to CVPR!
  • [Talk] Carl Vondrick from Google gave a talk as part of the CLVR Speaker Series.
  • [Paper] Paper on agents that disentangle environment and task accepted to ICLR!
  • [Talk] Jimmy Ba from UToronto gave a talk as part of the CLVR Speaker Series.
  • [Talk] Phillip Isola from OpenAI gave a talk as part of the CLVR Speaker Series.
2017
  • [Paper] Paper on multi-modal imitation learning accepted to NIPS!