Hang Yin

Описание: Talking Robotics #2

Speaker
Hang Yin, KTH

Title
Efficient Representations in Learning Visual Planning and Contact-rich Tasks

Speaker Bio
Hang is a Postdoctoral Researcher at the Robotics, Perception and Learning Group, KTH Royal Institute of Technology. Hang’s interests lie in the intersection of robotics and machine learning and he is enthusiastic about finding and integrating problem structures, such as task representation, dynamical systems and optimization-based control, to facilitate learning-based robotics.

Hang obtained his PhD from EPFL and IST, University of Lisbon under the supervision of Prof. Aude Billard, Prof. Ana Paiva and Prof. Francisco S. Melo. Prior to that, Hang completed his master and bachelor studies in Shanghai Jiao Tong University. He also worked as a software engineer in Siemens.

Speaker Links
Website: https://navigator8972.github.io
GitHub: https://github.com/navigator8972
Google Scholar: https://scholar.google.pt/citations?u...

Abstract
Exploiting efficient representations is the key to learning and automating real-world robotic tasks. In this talk, I will present recent results at RPL about learning state or modeling action spaces for robotic agents to address challenging manipulation tasks. In the first part, I will present a framework of learning a compact representation to encode high-dimensional and complex state/observations, e.g. configurations of a clothing item. I will show a variant of variational autoencoders which imposes a low-dimensional latent space with an improved structure. This allows us to build a roadmap in the embedding space and perform effective action planning in a cloth folding task. In the second part, the talk will focus on incorporating well-established robot control results in the reinforcement learning action design. I will introduce a policy in the form of a stable variable impedance controller and a variant of Cross Entropy Method to guarantee stable explorations in learning contact-rich skills. Our results demonstrate a superior performance in simulated and real-world peg-in-hole tasks, despite admitting a more restricted class of behaviors compared to baseline neural network policies.

Papers covered during the talk
Lippi, M., Poklukar, P., Welle, M. C., Varava, A., Yin, H., Marino, A., & Kragic, D. (2020). Latent Space Roadmap for Visual Action Planning of Deformable and Rigid Object Manipulation. arXiv preprint arXiv:2003.08974.
Khader, S. A., Yin, H., Falco, P., & Kragic, D. (2020). Stability-Guaranteed Reinforcement Learning for Contact-rich Manipulation. arXiv preprint arXiv:2004.10886.

Chapters
00:00 Presenter Introduction
01:40 Presentation
27:57 Q&A
48:28 End

/////////////////////////////////////////////

Talking Robotics is a series of virtual seminars about Robotics and its interaction with other relevant fields, such as Artificial Intelligence, Machine Learning, Design Research, Human-Robot Interaction, among others. We aim to promote reflections, dialogues, and a place to network.

Organized by Patrícia Alves-Oliveira, Silvia Tulli, Miguel Vasco, and Joana Campos

Email: talkingrobotics at gmail dot com
Web Page: https://talking-robotics.github.io
Twitter: @talkingrobotics