Leveraging Habitat: A Platform for Embodied AI Research
Wednesday, September 16, 2020 (Central Time) — 3:30PM - 4:45PM
Within the computer vision community, there has been a shift from so called internet AI like MNIST, ImageNet, and other static data sets to embodied AI where agents can act and perceive within their environments. Embodied agents are a stepping stone towards deployment of AI agents in the real world, as they require agents to perceive, analyze, and react within 3D environments. Embodied agents present unique challenges, such as coupling perception with control, that are not present in static data sets. We provide an introduction to Habitat, a platform created by Facebook AI Research that allows for configurable agents and environments for the purpose of accelerating and creating benchmarks for embodied AI research. The primary objective of this workshop is to provide an introduction to Habitat for embodied AI research and encourage adoption of the platform by the research community. Our target audience is the embodied AI research community, which encompasses a number of fields such as computer vision, natural language processing, and robotics. We also target undergraduate and graduate students who aim to become more involved in this area of research.
Michael Piseno, Ali Bharwani