Theoretical Understanding of the Information Flow on Continual Learning Performance

Abstract

Continual Learning (CL) has generated attention as a method of avoiding Catastrophic Forgetting (CF) in the sequential training of neural networks, improving network efficiency and adaptability to different tasks. Additionally, CL serves as an ideal setting for studying network behavior and Forward Knowledge Transfer (FKT) between tasks. Pruning methods for CL train subnetworks to handle the sequential tasks which allows us to take a structured approach to investigating FKT. Sharing prior subnetworks' weights leverages past knowledge for the current task through FKT. Understanding which weights to share is important as sharing all weights can yield sub-optimal accuracy. This paper investigates how different sharing decisions affect the FKT between tasks. Through this lens we demonstrate how task complexity and similarity influence the optimal weight sharing decisions, giving insights into the relationships between tasks and helping inform decision making in similar CL methods. We implement three sequential datasets designed to emphasize variation in task complexity and similarity, reporting results for both ResNet-18 and VGG-16. By sharing in accordance with the decisions supported by our findings, we show that we can improve task accuracy compared to other sharing decisions.

Similarity-Based Adaptation for Task-Aware and Task-Free Continual Learning

“…Other previous works which touched on adaptive CL include (Nguyen et al, 2019;Ramasesh et al, 2021;Doan et al, 2021;Andle et al, 2023;Lin et al, 2023). Veniat et al (2021) used modules and leveraged a task-driven prior over the exponential search space to decide which modules to reuse.…”