Active Learning for Abstractive Text Summarization via LLM-Determined Curriculum and Certainty Gain Maximization
Dongyuan Li, Ying Zhang, Zhen Wang, Shiyin Tan, Satoshi Kosugi, Manabu Okumura · 2024
For abstractive text summarization (ATS), laborious data annotation and time-consuming model training become two high walls, hindering its further progress.Active Learning (AL), selecting a few informative instances for annotation and model training, sheds light on solving these issues.However, only few AL-based studies focus on ATS and suffer from low stability, effectiveness, and efficiency.To solve the problems, we propose a novel LLM-determined curriculum active learning framework.Firstly, we design a prompt to ask large language models to rate the difficulty of instances, which guides the model to train on from easier to harder instances.Secondly, we design a novel AL strategy, i.e., Certainty Gain Maximization, enabling to select instances whose distribution aligns well with the overall distribution.Experiments show that our method can improve the stability, effectiveness, and efficiency of the ATS backbones.Code is available on Github 1 .