Unsupervised Progressive Learning and the STAM Architecture
This addresses the problem of learning from streaming, unlabeled data without replay for AI systems, but it is incremental as it builds on existing continual learning methods.
The paper tackles the Unsupervised Progressive Learning (UPL) problem, which involves online representation learning from a non-stationary, unlabeled data stream without storing data, and proposes the Self-Taught Associative Memory (STAM) architecture to learn persistent features, achieving results comparable to adapted continual learning models like MAS and GEM in clustering and classification tasks.
We first pose the Unsupervised Progressive Learning (UPL) problem: an online representation learning problem in which the learner observes a non-stationary and unlabeled data stream, learning a growing number of features that persist over time even though the data is not stored or replayed. To solve the UPL problem we propose the Self-Taught Associative Memory (STAM) architecture. Layered hierarchies of STAM modules learn based on a combination of online clustering, novelty detection, forgetting outliers, and storing only prototypical features rather than specific examples. We evaluate STAM representations using clustering and classification tasks. While there are no existing learning scenarios that are directly comparable to UPL, we compare the STAM architecture with two recent continual learning models, Memory Aware Synapses (MAS) and Gradient Episodic Memories (GEM), after adapting them in the UPL setting.