Recent Work

Learning Structured Dynamics Models	Learning Composable Robot Skills from Videos	Hallucinating 3D Shapes from single Images for Grasping

Composing Pick-And-Place Tasks By Grounding Language		Learning Object Placements by Hallucinating Scenes

News

• 30 Nov 2021: Our work on leveraging visual affordances for efficient policy learning has been accepted for presentation at the NeurIPS 2021 workshop on Robot Learning.
• 15 Mar 2021: I am happy to co-organize an ICRA workshop on Semantic Representations for Robotics through Continuous Interaction and Incremental Learning together with Tamim Asfour and Wolfram Burgard.
• 17 Feb 2021: New preprint for our upcoming ISER paper on Composing Pick-And-Place Tasks By Grounding Language.
• 04 Aug 2020: New preprint for our upcoming IROS paper on Hindsight for Foresight: Unsupervised Structured Dynamics Models from Physical Interaction, check out the video.
• 13 May 2020: Our paper on unsupervised skill learning from video has been nominated for the Best Paper Award in Cognitive Robotics at ICRA 2020!
• 09 Mar 2020: I am co-organizing an RSS workshop on Self-supervised Robot Learning together with Wolfram Burgard, Anelia Angelova, Joschka Boedecker and Abhinav Valada. The submission deadline will be in April 2020!
• 12 Feb 2020: Our work on Learning Object Placements For Relational Instructions by Hallucinating Scene Representations has been featured on TechXplore.
• 01 Feb 2020: Two new papers accepted for ICRA 2020. Adversarial Skill Networks: Unsupervised Robot Skill Learning from Video and Learning Object Placements For Relational Instructionsby Hallucinating Scene Representations.