Embodied Question Answering

TL;DR

This paper introduces Embodied Question Answering, a complex AI task where agents navigate 3D environments to answer questions, integrating perception, navigation, and reasoning skills.

Contribution

It defines the EmbodiedQA task, develops environments and evaluation protocols, and trains reinforcement learning agents to perform this integrated task.

Findings

Agents can navigate and gather information in 3D environments.

Reinforcement learning enables agents to answer questions based on visual exploration.

The task combines perception, navigation, and reasoning skills.

Abstract

We present a new AI task -- Embodied Question Answering (EmbodiedQA) -- where an agent is spawned at a random location in a 3D environment and asked a question ("What color is the car?"). In order to answer, the agent must first intelligently navigate to explore the environment, gather information through first-person (egocentric) vision, and then answer the question ("orange"). This challenging task requires a range of AI skills -- active perception, language understanding, goal-driven navigation, commonsense reasoning, and grounding of language into actions. In this work, we develop the environments, end-to-end-trained reinforcement learning agents, and evaluation protocols for EmbodiedQA.