Target Languages (vs. Inductive Biases) for Learning to Act and Plan
This addresses the challenge of generalization in AI for researchers and practitioners, offering a novel perspective but is incremental as it builds on existing symbolic AI ideas.
The paper tackles the problem of limited out-of-distribution generalization in deep learning and reinforcement learning by proposing a learning approach where representations are learned over a target language with known semantics, rather than relying on vague inductive biases. It illustrates this in the context of learning to act and plan, aiming to bridge symbolic and neural methods.
Recent breakthroughs in AI have shown the remarkable power of deep learning and deep reinforcement learning. These developments, however, have been tied to specific tasks, and progress in out-of-distribution generalization has been limited. While it is assumed that these limitations can be overcome by incorporating suitable inductive biases, the notion of inductive biases itself is often left vague and does not provide meaningful guidance. In the paper, I articulate a different learning approach where representations do not emerge from biases in a neural architecture but are learned over a given target language with a known semantics. The basic ideas are implicit in mainstream AI where representations have been encoded in languages ranging from fragments of first-order logic to probabilistic structural causal models. The challenge is to learn from data the representations that have traditionally been crafted by hand. Generalization is then a result of the semantics of the language. The goals of this paper are to make these ideas explicit, to place them in a broader context where the design of the target language is crucial, and to illustrate them in the context of learning to act and plan. For this, after a general discussion, I consider learning representations of actions, general policies, and subgoals ("intrinsic rewards"). In these cases, learning is formulated as a combinatorial problem but nothing prevents the use of deep learning techniques instead. Indeed, learning representations over languages with a known semantics provides an account of what is to be learned, while learning representations with neural nets provides a complementary account of how representations can be learned. The challenge and the opportunity is to bring the two together.