A Probabilistic Framework for Modular Continual Learning

arXiv:2306.06545v2 Announce Type: replace
Abstract: Modular approaches that use a different composition of modules for each problem are a promising direction in continual learning (CL). However, searching through the large, discrete space of module compositions is challenging, especially because evaluating a composition's performance requires a round of neural network training. We address this challenge through a modular CL framework, PICLE, that uses a probabilistic model to cheaply compute the fitness of each composition, allowing PICLE to achieve both perceptual, few-shot …

abstract arxiv continual cs.lg framework however modular modules network network training neural network performance searching space s performance stat.ml through training type