Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration
It addresses the problem of developing socially intelligent AI agents for human-AI collaboration, but is incremental as it builds on existing virtual environment and benchmark methods.
The paper introduces the Watch-And-Help challenge to test social intelligence in AI agents by requiring them to understand tasks from demonstrations and collaborate with human-like agents in household environments, with results showing it enables systematic evaluation of machine social intelligence.
In this paper, we introduce Watch-And-Help (WAH), a challenge for testing social intelligence in agents. In WAH, an AI agent needs to help a human-like agent perform a complex household task efficiently. To succeed, the AI agent needs to i) understand the underlying goal of the task by watching a single demonstration of the human-like agent performing the same task (social perception), and ii) coordinate with the human-like agent to solve the task in an unseen environment as fast as possible (human-AI collaboration). For this challenge, we build VirtualHome-Social, a multi-agent household environment, and provide a benchmark including both planning and learning based baselines. We evaluate the performance of AI agents with the human-like agent as well as with real humans using objective metrics and subjective user ratings. Experimental results demonstrate that the proposed challenge and virtual environment enable a systematic evaluation on the important aspects of machine social intelligence at scale.