ICML 2019 Expo Demo
July 12, 2020
Reaching intent estimation via approximate Bayesian computation
Sponsor: Intel AI
Javier Felip Leon (Speaker)
This demo shows a system that provides real-time user intent estimation. The environment consists of a table and a user in front of it, and the task consists of the user placing a fiducial object on the table. The system estimates the intended placement location and represents it as a probability density function over the table. The system is composed of three elements, an object tracker, a model-based physically plausible trajectory generator and a probability function. The user is captured through a Realsense RGBD camera and the intent is obtained through approximate Bayesian computation in an analysis by synthesis approach. The demo is interactive and users are allowed to play with it and observe the behavior of the predictions (see video linked). All the computations, including tracking, are performed in a NUC.