Continual Learning with Pre-Trained Models: A Survey
It addresses the challenge of catastrophic forgetting in streaming data applications for researchers and practitioners, but it is incremental as it surveys and analyzes existing work rather than introducing new methods.
This survey tackles the problem of continual learning with pre-trained models by categorizing existing methods and providing a comparative analysis, including an empirical study to highlight fairness concerns in evaluations.
Nowadays, real-world applications often face streaming data, which requires the learning system to absorb new knowledge as data evolves. Continual Learning (CL) aims to achieve this goal and meanwhile overcome the catastrophic forgetting of former knowledge when learning new ones. Typical CL methods build the model from scratch to grow with incoming data. However, the advent of the pre-trained model (PTM) era has sparked immense research interest, particularly in leveraging PTMs' robust representational capabilities. This paper presents a comprehensive survey of the latest advancements in PTM-based CL. We categorize existing methodologies into three distinct groups, providing a comparative analysis of their similarities, differences, and respective advantages and disadvantages. Additionally, we offer an empirical study contrasting various state-of-the-art methods to highlight concerns regarding fairness in comparisons. The source code to reproduce these evaluations is available at: https://github.com/sun-hailong/LAMDA-PILOT