To Softmax, or not to Softmax: that is the question when applying Active Learning for Transformer Models
This addresses the challenge of reducing labeled data needs for fine-tuning Transformers in NLP, but it is incremental as it focuses on improving confidence measures within existing active learning frameworks.
The paper tackled the problem of using misleading softmax probabilities for active learning in Transformer models, finding that labeling only the most uncertain samples (outliers) worsens performance, and proposed a heuristic to ignore samples systematically, improving methods over softmax on seven datasets.
Despite achieving state-of-the-art results in nearly all Natural Language Processing applications, fine-tuning Transformer-based language models still requires a significant amount of labeled data to work. A well known technique to reduce the amount of human effort in acquiring a labeled dataset is \textit{Active Learning} (AL): an iterative process in which only the minimal amount of samples is labeled. AL strategies require access to a quantified confidence measure of the model predictions. A common choice is the softmax activation function for the final layer. As the softmax function provides misleading probabilities, this paper compares eight alternatives on seven datasets. Our almost paradoxical finding is that most of the methods are too good at identifying the true most uncertain samples (outliers), and that labeling therefore exclusively outliers results in worse performance. As a heuristic we propose to systematically ignore samples, which results in improvements of various methods compared to the softmax function.