Workshop Summary

Advances in robot learning in recent years have yielded outstanding performance across robotic tasks, compelling roboticists to reexamine the role of long-standing, reliable probabilistic inference algorithms. Probabilistic methods have a long history of enabling effective operation under uncertain and unstructured environments and provide a unifying perspective on perception, control and learning. Robot learning, on the other hand, promises generalizability, eliminating the need for carefully hand-crafted models required in many classical probabilistic algorithms. However, it remains unclear whether we can achieve reliable, adaptable behavior by relying on data alone. This workshop sets out to address the question: How can roboticists achieve the best of both deep learning and probabilistic inference?

Recently, the robotics community has seen an emergence of interest in hybrid methods which aim to exploit the benefits of both underlying model classes. Key developments in Differentiable Probabilistic Robotics have included the introduction of learned components within inference frameworks and end-to-end differentiable algorithms for Bayesian inference. These methods offer the opportunity to develop robust and reliable learning and adaptive systems. This workshop aims to connect researchers working at the intersection of robotics, deep learning, and probabilistic inference to facilitate breakthrough research in these areas.

Announcements

  • We are excited to announce that the workshop has over 150 registrants as of September 18.
  • Selected best papers will receive monetary prizes (sponsored by RAS TC for Computer & Robot Vision).
  • Our workshop will take place on Sunday, October 1, 2023 in
    Room 330A, Huntington Place, Detroit, MI. USA
  • Submissions are now open! Please see the Call for Contributions for details.
    Due August 25, 2023 August 30, 2023.

Workshop Photos

Program

The following program is tentative.

8:45 – 9:00 Welcome message from organizers
9:00 – 9:40 Rudolph Triebel and Jongseok Lee
Seeing the Unknown - Probabilistic Reasoning for Introspective Robot Perception
9:40 – 10:20 Nicholas Roy
The Role of Probability in Differential Learned Models
10:20 – 10:50 Poster session & Coffee break
10:50 – 11:30 Rika Antonova
Navigating Differentiable Simulation Landscapes with Bayesian Optimization and Continual Learning
11:30 – 12:10 Hanna Kurniawati
POMDP and Learning: Utilising Primitive Computation Information
12:10 – 13:10 Lunch break
13:10 – 13:50 Georgia Chalvatzaki
Batching Everything Everywhere for Fast Robot Planning and Control
13:50 – 14:30 Florian Shkurti
Beyond System Identification: Differentiable Physics and Rendering for Prediction, Safety, and Task and Motion Planning
14:30 – 14:45 Spotlight Talk 1
Probabilistic Differentiable Filters Enable Ubiquitous Robot Control with Smartwatches
14:45 – 15:00 Spotlight Talk 2
Learning from Demonstration via Probabilistic Diagrammatic Teaching
15:00 – 15:50 Poster session & Coffee break
15:50 – 16:30 Peter Karkus
Differentiable Robotics: Compositional Deep Learning with Differentiable Algorithm Networks
16:30 – 17:15 Panel Discussion
Hanna Kurniawati, Peter Karkus, Rika Antonova, Rudolph Triebel, Florian Shkurti, and Tucker Hermans
17:15 – 17:30 Closing remarks from organizers

Accepted Papers

Accelerating Motion Planning via Optimal Transport An Thai Le, Georgia Chalvatzaki, Armin Biess, Jan Peters
Probabilistic Differentiable Filters Enable Ubiquitous Robot Control with Smartwatches Fabian C Weigend, Xiao Liu, Heni Ben Amor
Probabilistic Pick and Place Planning instead of Pick then Place Planning Mohanraj Devendran Shanthi, Tucker Hermans
Learning from Demonstration via Probabilistic Diagrammatic Teaching Weiming Zhi, Tianyi Zhang, Matthew Johnson-Roberson
Sampling Constrained Trajectories Using Composable Diffusion Models Thomas Power, Rana Soltani-Zarrin, Soshi Iba, Dmitry Berenson
Task-space Kernels for Diverse Stein Variational MPC Madhav Shekhar Sharma, Thomas Power, Dmitry Berenson
Adaptive Magnetic Control using Stein Variational Gradient Descent computed Distribution of Object Parameters Griffin Tabor, Tucker Hermans
Cooperative Probabilistic Trajectory Forecasting Under Occlusion Anshul Nayak, Azim Eskandarian

Sponsors

Endorsements

This workshop is proud to be endorsed by the IEEE RAS Technical Committees for Computer and Robot Vision, Robot Learning, and Algorithms for Planning and Control of Robot Motion.