Visual Learning and Reasoning for Robotic Manipulation

Full-day workshop at RSS 2020

July 13, 2020

Oregon State University at Corvallis, Oregon, USA

Update on the Coronavirus:

  • March 29: Due to concerns about COVID-19, RSS 2020 will take place VIRTUALLY. Accordingly, our workshop will hold presentations and discussions in a virtual format on the original date (July 13, 2020). Each contributed paper will be presented remotely in a spotlight talk session. The paper submission deadline has been extended to May 4, 2020. We will keep you informed as further decisions are made.

Abstract

Visual perception is essential for achieving robot autonomy in the real world. To perform complex manipulation tasks in unknown environments, a robot needs to actively acquire knowledge through physical interactions and conduct sophisticated reasoning of the observed objects. This invites a series of research challenges in developing computational tools to close the perception-action loop. Given the recent advances in computer vision and deep learning, we look for new potential solutions for performing real-world robotic manipulation in an effective and computationally efficient manner. We focus on the two parallel themes in this workshop:

  • How could a robot’s interaction with the physical world facilitate the development of its visual perception?
  • How a deep understanding of the physical world through visual learning and reasoning could give rise to effective and robust robotic control?


The invited talks and contributed papers will incorporate a variety of relevant subtopics. This includes, but is not limited to, the following topics:

  • Image-based robotic manipulation and grasping
  • Visual representation learning from interaction
  • Learning algorithms for visuomotor skills from raw sensory data
  • Visual reasoning of object affordances
  • Active perception and physical scene understanding
  • Sim-to-real transfer and domain adaptation for visuomotor policies

Call for Papers

We're inviting submissions! If you're interested in (remotely) presenting a spotlight talk, please submit a short paper (or extended abstract) to CMT.

We suggest extended abstracts of 2 pages in the RSS format. A maximum of 4 pages will be considered. References will not count towards the page limit. Accepted papers will be listed on the workshop website.

Important Dates:

  • Paper Submission: May 4, 2020
  • Author Notification: May 25, 2020
  • Conference Date: July 13, 2020

Invited Speakers

Invited speakers from academia and industry will give talks in this workshop. Eight speakers have confirmed or tentatively confirmed:

University of Washington / Nvidia

Stanford University

University of Pennsylvania

Stanford University

Google Research

Georgia Tech

UC Berkeley

MIT / TRI

Tentative Schedule

9:00 am - 9:15 am: Opening remark

9:15am - 9:45am: Invited talk: Speaker 1

9:45am - 10:15am: Invited talk: Speaker 2

10:15pm - 10:45pm: Spotlight talk session

10:45pm - 11:15pm: Invited talk: Speaker 3

11:15am - 11:45pm: Invited talk: Speaker 4

11:45am - 12:00pm: Poster teasers

1:30pm - 2:00pm: Invited talk: Speaker 5

2:00pm - 2:30pm: Invited talk: Speaker 6

3:00pm - 3:30pm: Spotlight talk session

3:30pm - 4:00pm: Invited talk: Speaker 7

4:00pm - 4:30pm: Invited talk: Speaker 8

4:30pm - 5:00pm: Panel discussion

Organizers

Stanford University

Carnegie Mellon University

UT Austin / Nvidia

University of Pennsylvania

University of Toronto / Nvidia

Samsung Research

Nvidia

McGill University / Samsung

Contact

Kuan Fang

kuanfang [AT] stanford [DOT] edu

Sponsor