Towards learning-to-learn
This foundational idea could impact all of AI by potentially replacing human-designed learning algorithms with learned ones, but it is currently conceptual and incremental in its discussion.
The paper discusses the principle of learning-to-learn, arguing that current AI systems rely on human-designed learning rules like gradient descent, which may be suboptimal, and proposes that learning itself should be learned to improve performance, drawing connections across machine learning, neuroscience, and cognitive science.
In good old-fashioned artificial intelligence (GOFAI), humans specified systems that solved problems. Much of the recent progress in AI has come from replacing human insights by learning. However, learning itself is still usually built by humans -- specifically the choice that parameter updates should follow the gradient of a cost function. Yet, in analogy with GOFAI, there is no reason to believe that humans are particularly good at defining such learning systems: we may expect learning itself to be better if we learn it. Recent research in machine learning has started to realize the benefits of that strategy. We should thus expect this to be relevant for neuroscience: how could the correct learning rules be acquired? Indeed, cognitive science has long shown that humans learn-to-learn, which is potentially responsible for their impressive learning abilities. Here we discuss ideas across machine learning, neuroscience, and cognitive science that matter for the principle of learning-to-learn.