Overview

The computer vision and robotics communities have recently seen tremendous success in powerful, accurate visual reconstructions via implicit scene representations (neural radiance fields, signed distance functions, etc.). Similarly, there have been amazing successes in implicit policy representations for complex tasks. However, it remains unclear how to create or use such implicit representations within robotics for arbitrary manipulation tasks.

Given the exciting, rapidly-evolving state of the art in implicit representations, this workshop seeks to bring together a variety of speakers and participants in robotic manipulation, robot learning, planning, and computer vision to explore which role implicit representations can play in solving robotics tasks.

Talks

• Dieter Fox [Recorded Talk]
  University of Washington, NVIDIA
• Danica Kragic [Recorded Talk]
  KTH
• Yunzhu Li, Danny Driess [Recorded Talk]
  MIT, TU Berlin
• Jiajun Wu
  Stanford University
• Animesh Garg [Recorded Talk]
  University of Toronto
• Igor Mordatch [Recorded Talk]
  Google Brain
• Vincent Sitzmann [Recorded Talk]
  MIT
• Andy Zeng [Recorded Talk]
  Google Robotics

Accepted Contributions

• Learning Implicit Priors for Motion Optimization
  Alexander Lambert, An T. Le, Julen Urain, Georgia Chalvatzaki, Byron Boots, Jan Peters
• Conditional Energy-Based Models for Implicit Policies: The Gap between Theory and Practice
  Duy-Nguyen Ta, Eric Cousineau, Huihua Zhao, Siyuan Feng
• Representations for Scheduling of Heterogeneous Computation to Support Motion Planning
  Justin McGowen, Ismet Dagli, Mehmet Belviranli, Neil Dantam
• NeRF-Frenemy: Co-Opting Adversarial Learning for Autonomy-Directed Co-Design
  Stanley Lewis, Bahaa Aldeeb, Anthony Opipari, Elizabeth A. Olson, Cameron Kisailus, Odest Chadwicke Jenkins
• Active View Planning for Radiance Fields
  Kevin Lin, Brent Yi
• Deep Visual Constraints: Neural Implicit Models for Manipulation Planning from Visual Input
  Jung-Su Ha, Danny Driess, Marc Toussaint
• Integrating Symmetry into Differentiable Planning
  Linfeng Zhao, Xupeng Zhu, Lingzhi Kong, Robin Walters, Lawson L.S. Wong
• Neural Motion Fields: Encoding Grasp Trajectories as Implicit Value Functions
  Yun-Chun Chen, Adithyavairavan Murali, Balakumar Sundaralingam, Wei Yang, Animesh Garg, Dieter Fox
• Affordance Transfer based on Self-Aligning Implicit Representations of Local Surfaces
  Ahmet Tekden, Marc Peter Deisenroth, Yasemin Bekiroglu
• Implicit Object Mapping With Noisy Data
  Jad Abou-Chakra, Feras Dayoub, Niko Sunderhauf
• Learning Explicit Infeasibility from Implicit Configuration Space Connectivity
  Sihui Li, Neil T. Dantam
• Robot Skill Adaptation via Soft Actor-Critic Gaussian Mixture Models
  Iman Nematollahi, Erick Rosete-Beas, Adrian Röfer, Tim Welschehold, Abhinav Valada, Wolfram Burgard

Organizers

• Danny Driess
  Technical University of Berlin
• Yunzhu Li
  Massachusetts Institute of Technology
• Shreyas Kousik
  Stanford University
• Jeannette Bohg
  Stanford University
• Russ Tedrake
  Massachusetts Institute of Technology
• Marc Toussaint
  Technical University of Berlin

Schedule