IGLU Gridworld: Simple and Fast Environment for Embodied Dialog Agents
Artem Zholus, Alexey Skrynnik, Shrestha Mohanty, Zoya Volovikova, Julia Kiseleva, Artur Szlam, Marc-Alexandre Coté, Aleksandr I. Panov
MicrosoftMIT
arXiv:2206.00142v113 citationsh-index: 26
Originality Synthesis-oriented
AI Analysis
This provides a tool for researchers in embodied AI and reinforcement learning to develop agents that understand and act on language instructions in complex environments.
The authors tackled the need for a scalable environment to train and evaluate language-conditioned embodied agents, resulting in the IGLU Gridworld, which includes visual embodiment, interactive collaboration, and hard 3D block-building tasks.
We present the IGLU Gridworld: a reinforcement learning environment for building and evaluating language conditioned embodied agents in a scalable way. The environment features visual agent embodiment, interactive learning through collaboration, language conditioned RL, and combinatorically hard task (3d blocks building) space.