Learning to Perform Complex Tasks through Compositional Fine-Tuning of Language Models
This addresses the problem of dependency on huge pretrained models for compositional tasks, offering a more efficient and data-flexible approach for domains like recommendation and inferential tasks.
The paper tackles the challenge of encoding compositional task structure in AI by introducing compositional fine-tuning (CFT), which decomposes tasks into components and fine-tunes smaller language models on a curriculum, resulting in outperforming end-to-end learning with equal data and matching chain of thought prompting using models only 7.4% the size.
How to usefully encode compositional task structure has long been a core challenge in AI. Recent work in chain of thought prompting has shown that for very large neural language models (LMs), explicitly demonstrating the inferential steps involved in a target task may improve performance over end-to-end learning that focuses on the target task alone. However, chain of thought prompting has significant limitations due to its dependency on huge pretrained LMs. In this work, we present compositional fine-tuning (CFT): an approach based on explicitly decomposing a target task into component tasks, and then fine-tuning smaller LMs on a curriculum of such component tasks. We apply CFT to recommendation tasks in two domains, world travel and local dining, as well as a previously studied inferential task (sports understanding). We show that CFT outperforms end-to-end learning even with equal amounts of data, and gets consistently better as more component tasks are modeled via fine-tuning. Compared with chain of thought prompting, CFT performs at least as well using LMs only 7.4% of the size, and is moreover applicable to task domains for which data are not available during pretraining.