Iterative Decoding for Compositional Generalization in Transformers

Abstract

Deep learning models generalize well to in-distribution data but struggle to generalize compositionally, i.e., to combine a set of learned primitives to solve more complex tasks. In sequence-to-sequence (seq2seq) learning, transformers are often unable to predict correct outputs for longer examples than those seen at training. This paper introduces iterative decoding, an alternative to seq2seq that (i) improves transformer compositional generalization in the PCFG and Cartesian product datasets and (ii) evidences that, in these datasets, seq2seq transformers do not learn iterations that are not unrolled. In iterative decoding, training examples are broken down into a sequence of intermediate steps that the transformer learns iteratively. At inference time, the intermediate outputs are fed back to the transformer as intermediate inputs until an end-of-iteration token is predicted. We conclude by illustrating some limitations of iterative decoding in the CFQ dataset.

References in corpus (6)

• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
• Compositional Generalization in Semantic Parsing: Pre-training vs. Specialized Architectures
• HIBERT: Document Level Pre-training of Hierarchical Bidirectional Transformers for Document Summarization
• ETC: Encoding Long and Structured Inputs in Transformers
• PonderNet: Learning to Ponder
• Improving Compositional Generalization in Classification Tasks via Structure Annotations