Towards a Unified View of Parameter-Efficient Transfer Learning

TL;DR

This paper presents a unified framework for parameter-efficient transfer learning in NLP, clarifying the relationships among methods and enabling the development of more effective, less parameter-intensive fine-tuning approaches.

Contribution

It introduces a unified view that models various parameter-efficient methods as modifications to hidden states, facilitating understanding and transfer of design principles across approaches.

Findings

Unified framework clarifies connections among methods

New methods tune fewer parameters with comparable performance

Empirical validation across multiple NLP tasks

Abstract

Fine-tuning large pre-trained language models on downstream tasks has become the de-facto learning paradigm in NLP. However, conventional approaches fine-tune all the parameters of the pre-trained model, which becomes prohibitive as the model size and the number of tasks grow. Recent work has proposed a variety of parameter-efficient transfer learning methods that only fine-tune a small number of (extra) parameters to attain strong performance. While effective, the critical ingredients for success and the connections among the various methods are poorly understood. In this paper, we break down the design of state-of-the-art parameter-efficient transfer learning methods and present a unified framework that establishes connections between them. Specifically, we re-frame them as modifications to specific hidden states in pre-trained models, and define a set of design dimensions along which different methods vary, such as the function to compute the modification and the position to apply the modification. Through comprehensive empirical studies across machine translation, text summarization, language understanding, and text classification benchmarks, we utilize the unified view to identify important design choices in previous methods. Furthermore, our unified framework enables the transfer of design elements across different approaches, and as a result we are able to instantiate new parameter-efficient fine-tuning methods that tune less parameters than previous methods while being more effective, achieving comparable results to fine-tuning all parameters on all four tasks.