Function-space Parameterization of Neural Networks for Sequential Learning
This addresses a key problem in continual learning and model-based reinforcement learning for AI systems that need to adapt over time, though it appears incremental as it builds on existing function-space methods.
The paper tackled the challenge of sequential learning in neural networks, which often struggles with incorporating new data and retaining prior knowledge, by introducing a function-space parameterization that enables efficient knowledge retention and data incorporation without retraining.
Sequential learning paradigms pose challenges for gradient-based deep learning due to difficulties incorporating new data and retaining prior knowledge. While Gaussian processes elegantly tackle these problems, they struggle with scalability and handling rich inputs, such as images. To address these issues, we introduce a technique that converts neural networks from weight space to function space, through a dual parameterization. Our parameterization offers: (i) a way to scale function-space methods to large data sets via sparsification, (ii) retention of prior knowledge when access to past data is limited, and (iii) a mechanism to incorporate new data without retraining. Our experiments demonstrate that we can retain knowledge in continual learning and incorporate new data efficiently. We further show its strengths in uncertainty quantification and guiding exploration in model-based RL. Further information and code is available on the project website.