Meta Continual Learning via Dynamic Programming
This work provides a theoretical foundation for meta continual learning, addressing challenges like catastrophic forgetting, though it appears incremental as it builds on existing methods with a new theoretical lens.
The authors tackled the lack of theoretical frameworks for analyzing generalization and catastrophic forgetting in meta continual learning by developing a dynamic programming approach, and their method achieved accuracy comparable to or better than state-of-the-art on benchmark datasets.
Meta continual learning algorithms seek to train a model when faced with similar tasks observed in a sequential manner. Despite promising methodological advancements, there is a lack of theoretical frameworks that enable analysis of learning challenges such as generalization and catastrophic forgetting. To that end, we develop a new theoretical approach for meta continual learning~(MCL) where we mathematically model the learning dynamics using dynamic programming, and we establish conditions of optimality for the MCL problem. Moreover, using the theoretical framework, we derive a new dynamic-programming-based MCL method that adopts stochastic-gradient-driven alternating optimization to balance generalization and catastrophic forgetting. We show that, on MCL benchmark data sets, our theoretically grounded method achieves accuracy better than or comparable to that of existing state-of-the-art methods.