Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context LearningDownload PDF

Published: 31 Oct 2022, Last Modified: 03 Jul 2024NeurIPS 2022 AcceptReaders: Everyone
Keywords: few-shot learning, in-context learning, parameter-efficient training
TL;DR: We propose a new recipe T-Few for parameter-efficient few-shot learning that outperforms GPT-3 in-context learning.
Abstract: Few-shot in-context learning (ICL) enables pre-trained language models to perform a previously-unseen task without any gradient-based training by feeding a small number of training examples as part of the input. ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made. Parameter-efficient fine-tuning (PEFT) (e.g. adapter modules, prompt tuning, sparse update methods, etc.) offers an alternative paradigm where a small set of parameters are trained to enable a model to perform the new task. In this paper, we rigorously compare few-shot ICL and PEFT and demonstrate that the latter offers better accuracy as well as dramatically lower computational costs. Along the way, we introduce a new PEFT method called (IA)^3 that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters. We also propose a simple recipe based on the T0 model called T-Few that can be applied to new tasks without task-specific tuning or modifications. We validate the effectiveness of T-Few on completely unseen tasks by applying it to the RAFT benchmark, attaining super-human performance for the first time and outperforming the state-of-the-art by 6% absolute. All of the code used in our experiments will be publicly available.
Supplementary Material: pdf
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 1 code implementation](https://www.catalyzex.com/paper/few-shot-parameter-efficient-fine-tuning-is/code)
11 Replies