Continual Deep Learning by Functional Regularisation of Memorable Past
This addresses the problem of life-long learning for AI systems by combining regularisation and memory-based methods, though it is incremental as it builds on existing functional-regularisation ideas.
The paper tackles catastrophic forgetting in continual deep learning by introducing a functional-regularisation approach that uses memorable past examples, achieving state-of-the-art performance on standard benchmarks.
Continually learning new skills is important for intelligent systems, yet standard deep learning methods suffer from catastrophic forgetting of the past. Recent works address this with weight regularisation. Functional regularisation, although computationally expensive, is expected to perform better, but rarely does so in practice. In this paper, we fix this issue by using a new functional-regularisation approach that utilises a few memorable past examples crucial to avoid forgetting. By using a Gaussian Process formulation of deep networks, our approach enables training in weight-space while identifying both the memorable past and a functional prior. Our method achieves state-of-the-art performance on standard benchmarks and opens a new direction for life-long learning where regularisation and memory-based methods are naturally combined.