Ex-Model: Continual Learning from a Stream of Trained Models

Abstract

Learning continually from non-stationary data streams is a challenging research topic of growing popularity in the last few years. Being able to learn, adapt, and generalize continually in an efficient, effective, and scalable way is fundamental for a sustainable development of Artificial Intelligent systems. However, an agent-centric view of continual learning requires learning directly from raw data, which limits the interaction between independent agents, the efficiency, and the privacy of current approaches. Instead, we argue that continual learning systems should exploit the availability of compressed information in the form of trained models. In this paper, we introduce and formalize a new paradigm named "Ex-Model Continual Learning" (ExML), where an agent learns from a sequence of previously trained models instead of raw data. We further contribute with three ex-model continual learning algorithms and an empirical setting comprising three datasets (MNIST, CIFAR-10 and CORe50), and eight scenarios, where the proposed algorithms are extensively tested. Finally, we highlight the peculiarities of the ex-model paradigm and we point out interesting future research directions.

Practical Recommendations for Replay-based Continual Learning Methods

“…Continual learning algorithm are evaluated by benchmarks: they specify how the stream of data is created by defining the originating dataset(s), the amount of samples, the criteria to split the data in different tasks or experiences [6] and so on. In literature, different benchmarks are used to evaluate results.…”

“…The second direction (Sec. 6) is related to the balancing of the memory buffer. In the literature the replay buffer is usually balanced to have an equal amount of samples of each past task or class.…”