no code implementations • 7 Jul 2021 • Junha Roh, Karthik Desingh, Ali Farhadi, Dieter Fox
Specifically, given a reconstructed 3D scene in the form of point clouds with 3D bounding boxes of potential object candidates, and a language utterance referring to a target object in the scene, our model successfully identifies the target object from a set of potential candidates.
1 code implementation • 8 Nov 2020 • Junha Roh, Christoforos Mavrogiannis, Rishabh Madan, Dieter Fox, Siddhartha S. Srinivasa
Our key insight is that the geometric structure of the intersection and the incentive of agents to move efficiently and avoid collisions (rationality) reduces the space of likely behaviors, effectively relaxing the problem of trajectory prediction.
no code implementations • 16 Oct 2019 • Junha Roh, Chris Paxton, Andrzej Pronobis, Ali Farhadi, Dieter Fox
Widespread adoption of self-driving cars will depend not only on their safety but largely on their ability to interact with human users.