0

Grounded Language Learning Fast and Slow

An embodied agent with dual-coding external memory can perform one-shot word learning in a 3D environment, integrating short-term and long-term knowledge, and showing potential for fast-mapping.

Year
2020
Venue
ICLR 2021 1
Authors
6
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2009.01719v4ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Recent work has shown that large text-based neural language models, trained with conventional supervised learning objectives, acquire a surprising propensity for few- and one-shot learning. Here, we show that an embodied agent situated in a simulated 3D world, and endowed with a novel dual-coding external memory, can exhibit similar one-shot word learning when trained with conventional reinforcement learning algorithms. After a single introduction to a novel object via continuous visual perception and a language prompt ("This is a dax"), the agent can re-identify the object and manipulate it as instructed ("Put the dax on the bed"). In doing so, it seamlessly integrates short-term, within-episode knowledge of the appropriate referent for the word "dax" with long-term lexical and motor knowledge acquired across episodes (i.e. "bed" and "putting"). We find that, under certain training conditions and with a particular memory writing mechanism, the agent's one-shot word-object binding generalizes to novel exemplars within the same ShapeNet category, and is effective in settings with unfamiliar numbers of objects. We further show how dual-coding memory can be exploited as a signal for intrinsic motivation, stimulating the agent to seek names for objects that may be useful for later executing instructions. Together, the results demonstrate that deep neural networks can exploit meta-learning, episodic memory and an explicitly multi-modal environment to account for 'fast-mapping', a fundamental pillar of human cognitive development and a potentially transformative capacity for agents that interact with human users.

Authors

6