Learning with Latent Language

TL;DR

This paper demonstrates that leveraging natural language as a structured parameter space during pretraining enhances the generality and efficiency of classifiers and control policies across various tasks.

Contribution

It introduces a method that uses natural language strings as a parameter space, improving learning efficiency without requiring language data during task-specific training.

Findings

Models with linguistic parameterization outperform non-linguistic models.

Pretraining with language structures benefits image classification, text editing, and reinforcement learning.

Language-based parameter space improves generalization and learning speed.

Abstract

The named concepts and compositional operators present in natural language provide a rich source of information about the kinds of abstractions humans use to navigate the world. Can this linguistic background knowledge improve the generality and efficiency of learned classifiers and control policies? This paper aims to show that using the space of natural language strings as a parameter space is an effective way to capture natural task structure. In a pretraining phase, we learn a language interpretation model that transforms inputs (e.g. images) into outputs (e.g. labels) given natural language descriptions. To learn a new concept (e.g. a classifier), we search directly in the space of descriptions to minimize the interpreter's loss on training examples. Crucially, our models do not require language data to learn these concepts: language is used only in pretraining to impose structure on subsequent learning. Results on image classification, text editing, and reinforcement learning show that, in all settings, models with a linguistic parameterization outperform those without.