Text Representation Distillation via Information Bottleneck Principle
This addresses the computational cost and accessibility issues of PLMs for practical applications, but it is incremental as it builds on existing distillation methods.
The paper tackles the problem of performance degradation when distilling large pre-trained language models into smaller ones for text representation, proposing IBKD based on the Information Bottleneck principle to preserve key information while reducing overfitting, and shows effectiveness on Semantic Textual Similarity and Dense Retrieval tasks.
Pre-trained language models (PLMs) have recently shown great success in text representation field. However, the high computational cost and high-dimensional representation of PLMs pose significant challenges for practical applications. To make models more accessible, an effective method is to distill large models into smaller representation models. In order to relieve the issue of performance degradation after distillation, we propose a novel Knowledge Distillation method called IBKD. This approach is motivated by the Information Bottleneck principle and aims to maximize the mutual information between the final representation of the teacher and student model, while simultaneously reducing the mutual information between the student model's representation and the input data. This enables the student model to preserve important learned information while avoiding unnecessary information, thus reducing the risk of over-fitting. Empirical studies on two main downstream applications of text representation (Semantic Textual Similarity and Dense Retrieval tasks) demonstrate the effectiveness of our proposed approach.