DALLMi: Domain Adaption for LLM-based Multi-label Classifier

TL;DR

DALLMi is a novel semi-supervised domain adaptation method for multi-label text classification using LLMs, effectively handling domain shifts with limited labeled data and improving performance over existing approaches.

Contribution

It introduces a new variation loss, MixUp regularization, and label-balanced sampling for domain adaptation in multi-label text classification with LLMs.

Findings

DALLMi outperforms unsupervised methods by 19.9% mAP.

DALLMi outperforms partial supervision by 52.2% mAP.

Effective handling of domain shifts with limited labeled data.

Abstract

Large language models (LLMs) increasingly serve as the backbone for classifying text associated with distinct domains and simultaneously several labels (classes). When encountering domain shifts, e.g., classifier of movie reviews from IMDb to Rotten Tomatoes, adapting such an LLM-based multi-label classifier is challenging due to incomplete label sets at the target domain and daunting training overhead. The existing domain adaptation methods address either image multi-label classifiers or text binary classifiers. In this paper, we design DALLMi, Domain Adaptation Large Language Model interpolator, a first-of-its-kind semi-supervised domain adaptation method for text data models based on LLMs, specifically BERT. The core of DALLMi is the novel variation loss and MixUp regularization, which jointly leverage the limited positively labeled and large quantity of unlabeled text and, importantly, their interpolation from the BERT word embeddings. DALLMi also introduces a label-balanced sampling strategy to overcome the imbalance between labeled and unlabeled data. We evaluate DALLMi against the partial-supervised and unsupervised approach on three datasets under different scenarios of label availability for the target domain. Our results show that DALLMi achieves higher mAP than unsupervised and partially-supervised approaches by 19.9% and 52.2%, respectively.