Challenges

We plan to feature the following two challenges in the domain of multi-agent relational reasoning:

  • TrajNet++, a multi-agent trajectory forecasting challenge in the human behavior/motion domain

  • CLEVRER, a visual object dynamics reasoning challenge in the physical interaction domain

  • INTERPRET, an INTERACTION-Dataset-Based prediction challenge in the autonomous driving domain

There will be prizes for the challenge winners!

More details will come soon!