Active Few-Shot Learning with FASL
This work addresses the problem of quickly developing and deploying NLP models for real-world business needs with minimal data, though it appears incremental as it combines existing techniques.
The paper tackles the challenge of training text classification models with limited data by combining few-shot and active learning into FASL, a platform that enables iterative and fast model training, and investigates optimal active learning methods and a stopping criterion for annotation.
Recent advances in natural language processing (NLP) have led to strong text classification models for many tasks. However, still often thousands of examples are needed to train models with good quality. This makes it challenging to quickly develop and deploy new models for real world problems and business needs. Few-shot learning and active learning are two lines of research, aimed at tackling this problem. In this work, we combine both lines into FASL, a platform that allows training text classification models using an iterative and fast process. We investigate which active learning methods work best in our few-shot setup. Additionally, we develop a model to predict when to stop annotating. This is relevant as in a few-shot setup we do not have access to a large validation set.