Efficient Continual Learning with Modular Networks and Task-Driven Priors
This work aims to improve the efficiency and scalability of continual learning systems for researchers and practitioners by addressing limitations beyond catastrophic forgetting.
This paper addresses catastrophic forgetting in Continual Learning (CL) and introduces new benchmarks to evaluate knowledge transfer and sub-linear scaling of memory and compute. They propose a modular network architecture that reuses and instantiates modules as atomic skills, achieving competitive performance on existing CL benchmarks and superior results on their new, more challenging benchmarks.
Existing literature in Continual Learning (CL) has focused on overcoming catastrophic forgetting, the inability of the learner to recall how to perform tasks observed in the past. There are however other desirable properties of a CL system, such as the ability to transfer knowledge from previous tasks and to scale memory and compute sub-linearly with the number of tasks. Since most current benchmarks focus only on forgetting using short streams of tasks, we first propose a new suite of benchmarks to probe CL algorithms across these new axes. Finally, we introduce a new modular architecture, whose modules represent atomic skills that can be composed to perform a certain task. Learning a task reduces to figuring out which past modules to re-use, and which new modules to instantiate to solve the current task. Our learning algorithm leverages a task-driven prior over the exponential search space of all possible ways to combine modules, enabling efficient learning on long streams of tasks. Our experiments show that this modular architecture and learning algorithm perform competitively on widely used CL benchmarks while yielding superior performance on the more challenging benchmarks we introduce in this work.