Continual Learning as Computationally Constrained Reinforcement Learning
This addresses the long-standing problem of building efficient lifelong learning agents for advancing AI capabilities, but it is incremental as it focuses on clarification and formalization rather than new breakthroughs.
The paper tackles the challenge of designing agents that accumulate knowledge over a lifetime by formalizing continual learning concepts, introducing a framework and tools to stimulate research.
An agent that efficiently accumulates knowledge to develop increasingly sophisticated skills over a long lifetime could advance the frontier of artificial intelligence capabilities. The design of such agents, which remains a long-standing challenge of artificial intelligence, is addressed by the subject of continual learning. This monograph clarifies and formalizes concepts of continual learning, introducing a framework and set of tools to stimulate further research.