PIGLeT: Language Grounding Through Neuro-Symbolic Interaction in a 3D World

TL;DR

PIGLeT is a neuro-symbolic model that learns physical commonsense through interaction, grounding language in physical dynamics, and effectively predicting and describing physical events in a 3D environment.

Contribution

It introduces a novel neuro-symbolic approach combining physical dynamics and language models for grounded language understanding in 3D worlds.

Findings

Achieves over 80% accuracy in predicting physical outcomes from language.

Outperforms larger text-to-text models by over 10% in forecasting accuracy.

Produces human-judged more accurate natural language summaries of physical interactions.

Abstract

We propose PIGLeT: a model that learns physical commonsense knowledge through interaction, and then uses this knowledge to ground language. We factorize PIGLeT into a physical dynamics model, and a separate language model. Our dynamics model learns not just what objects are but also what they do: glass cups break when thrown, plastic ones don't. We then use it as the interface to our language model, giving us a unified model of linguistic form and grounded meaning. PIGLeT can read a sentence, simulate neurally what might happen next, and then communicate that result through a literal symbolic representation, or natural language. Experimental results show that our model effectively learns world dynamics, along with how to communicate them. It is able to correctly forecast "what happens next" given an English sentence over 80% of the time, outperforming a 100x larger, text-to-text approach by over 10%. Likewise, its natural language summaries of physical interactions are also judged by humans as more accurate than LM alternatives. We present comprehensive analysis showing room for future work.