TALM: Tool Augmented Language Models
This addresses the limitation of scale-dependent language models for tasks requiring real-time or private data access, offering a promising direction to enhance capabilities without relying solely on model size, though it appears incremental as it builds on existing tool-augmentation concepts.
The paper tackles the problem of language models lacking access to ephemeral or private data and APIs by introducing Tool Augmented Language Models (TALM), which combine text-only augmentation with non-differentiable tools and self-play bootstrapping, resulting in significant performance gains on knowledge-heavy QA and math tasks compared to non-augmented models at the same scale, including successful out-of-distribution inferences where baseline models fail.
Transformer based language models (LMs) demonstrate increasing performance with scale across a wide variety of tasks. Scale alone however cannot enable models to solve tasks that require access to ephemeral, changing, or private data that was unavailable at training time. Many useful tasks may also benefit from LMs being able to access APIs that read or modify state. In this work, we present Tool Augmented Language Models (TALM), combining a text-only approach to augment language models with non-differentiable tools, and an iterative "self-play" technique to bootstrap performance starting from few tool demonstrations. TALM exhibits strong performance on both a knowledge-heavy QA task and a reasoning oriented math task with simple tools. At a given model scale, TALM significantly outperforms non-augmented LMs. We further demonstrate that TALM successfully performs out-of-distribution inferences on both QA and math tasks, where non-augmented LMs fail. Our results suggest that Tool Augmented Language Models are a promising direction to enrich LMs' capabilities, with less dependence on scale.