Compositional generalization through meta sequence-to-sequence learning
This addresses a key limitation in neural networks for tasks requiring compositional reasoning, though it appears incremental as it builds on existing meta-learning and memory-augmented approaches.
The paper tackles the problem of compositional generalization in sequence-to-sequence neural networks, which often fail to combine new and existing concepts, and shows that meta sequence-to-sequence learning with memory-augmented networks can solve several SCAN tests for compositional learning and apply implicit rules to variables.
People can learn a new concept and use it compositionally, understanding how to "blicket twice" after learning how to "blicket." In contrast, powerful sequence-to-sequence (seq2seq) neural networks fail such tests of compositionality, especially when composing new concepts together with existing concepts. In this paper, I show how memory-augmented neural networks can be trained to generalize compositionally through meta seq2seq learning. In this approach, models train on a series of seq2seq problems to acquire the compositional skills needed to solve new seq2seq problems. Meta se2seq learning solves several of the SCAN tests for compositional learning and can learn to apply implicit rules to variables.