Compositional Generalization by Learning Analytical Expressions
This addresses a fundamental limitation in neural networks for AI systems that require human-like reasoning, though it is incremental as it builds on existing cognitive and neural approaches.
The paper tackles the problem of compositional generalization in neural networks by proposing a model that connects memory-augmented neural networks with analytical expressions, achieving 100% accuracy on the SCAN benchmark.
Compositional generalization is a basic and essential intellective capability of human beings, which allows us to recombine known parts readily. However, existing neural network based models have been proven to be extremely deficient in such a capability. Inspired by work in cognition which argues compositionality can be captured by variable slots with symbolic functions, we present a refreshing view that connects a memory-augmented neural model with analytical expressions, to achieve compositional generalization. Our model consists of two cooperative neural modules, Composer and Solver, fitting well with the cognitive argument while being able to be trained in an end-to-end manner via a hierarchical reinforcement learning algorithm. Experiments on the well-known benchmark SCAN demonstrate that our model seizes a great ability of compositional generalization, solving all challenges addressed by previous works with 100% accuracies.