VoxWorld is a multimodal simulation platform for situation- and context-aware intelligent systems. VoxWorld is situationally grounded in language, perception, and interaction, uses multimodal semantics to encode context in real-time, and provides rich contextual parsing and examinable environmental awareness, and accomodates both neural and symbolic approaches to creating interactive agents.
New Shared Task!
We will be presenting at the GENEA Workshop October 18, 2020
The following are some of the publications that have appeared from 2017-2020