5th Workshop on
Semantic Policy and Action Representations for Autonomous Robots (SPAR)
September 27, 2021 - Prague, Czech Republic
at IROS 2021
Thank you for attending the SPAR workshop!
NOTE: The SPAR workshop will be held via Zoom and GatherTown.
The SPAR workshop will be held between 14:00 and 21:15 CEST. Check out the below schedule.
The talks and the panel discussion will be held via Zoom or gCon:
Zoom link: https://chalmers.zoom.us/j/63911574176
Password: ------
The poster presentations of the contributed abstracts will be held via GatherTown:
Password: (same as zoom)
Schedule
Time Speaker
14:00 - 14:10 Welcome and introduction. Chris, Karinne, Jesse, Maru & Mehul
SESSION I: Semantic descriptions for robot execution
14:10 - 14:35 Tetsuya Ogata, Waseda University. "Toward Embodied Intelligence with Predictive Learning – From Data to Experiences"
14:35 - 15:00 Gayane Kazhoyan, University of Bremen. "Semantic Representations for Scalability and Transferability to Novel Domains in Mobile Manipulation"
15:00 - 15:25 Matteo Saveriano, University of Innsbruck. "Hierarchical action decomposition and motion learning for the execution of manipulation tasks"
15:25 - 15:50 Oliver Kroemer, Carnegie Mellon. "Learning to Structure Manipulation Skills"
15:50 - 16:00 Short Break
16:00 - 16:20 Spotlight presentations - Poster sessions - (7)
16:20 - 16:35 Short Break
SESSION II: Semantics in long-horizon tasks
16:35 - 17:00 George Konidaris, Brown University. "Signal to Symbol (via Skills)"
17:00 - 17:25 Karthik Desingh, Univ of Washington. "Learning Object-centric Representations for Robot Manipulation Tasks"
17:25 - 17:50 Aleksandra Faust, Google Brain. "Learning to Learn for RL"
17:50 - 18:15 Nakul Gopalan, Georgia Tech. "Learning Transferable Symbols and Language Groundings from Perceptual Data for Instruction Following"
18:15 - 18:25 Short Break
18:25 - 18:55 Panel discussion: George Konidaris, Aleksandra Faust, Oliver Kroemer
18:55 - 19:20 Coffee + gather town + posters
SESSION III: Semantics in HRI
19:20 - 19:45 Caelan Garrett. "Task and Motion Planning using Mixed Discrete, Continuous, Probabilistic, and Learned Representations"
19:45 - 20:10 Zhou Yu, Columbia University. "Teaching Machines through Natural Language Interactions"
20:10 - 20:35 Kalesha Bullard, DeepMind. "Towards Zero-Shot Emergent Communication for Embodied Agents"
20:35 - 20:40 Final remarks by the organizers