Meta-Learning to Compositionally Generalize
Preprint
Abstract
Natural language is compositional; the meaning of a sentence is a function of the meaning of its parts. This property allows humans to create and interpret novel sentences, generalizing robustly outside their prior experience. Neural networks have been shown to struggle with this kind of generalization, in particular performing poorly on tasks designed to assess compositional generalization (i.e. where training and testing distributions differ in ways that would be trivial for a compositional strategy to resolve). Their poor performance on these tasks may in part be due to the nature of supervised learning which assumes training and testing data to be drawn from the same distribution. We implement a meta-learning augmented version of supervised learning whose objective directly optimizes for out-of-distribution generalization. We construct pairs of tasks for meta-learning by sub-sampling existing training data. Each pair of tasks is constructed to contain relevant examples, as determined by a similarity metric, in an effort to inhibit models from memorizing their input. Experimental results on the COGS and SCAN datasets show that our similarity-driven meta-learning can improve generalization performance.Keywords
All Related Versions
This publication has 17 references indexed in Scilit:
- Span-based Semantic Parsing for Compositional Generalization2021
- Compositional Generalization and Natural Language Variation: Can a Semantic Parsing Approach Handle Both?2021
- Meta-Learning for Domain Generalization in Semantic Parsing2021
- COGS: A Compositional Generalization Challenge Based on Semantic Interpretation2020
- Compositionality and Generalization In Emergent Languages2020
- Good-Enough Compositional Data Augmentation2020
- RAT-SQL: Relation-Aware Schema Encoding and Linking for Text-to-SQL Parsers2020
- Visualisation and 'Diagnostic Classifiers' Reveal How Recurrent and Recursive Neural Networks Process Hierarchical StructureJournal of Artificial Intelligence Research, 2018
- Simple Fast Algorithms for the Editing Distance between Trees and Related ProblemsSIAM Journal on Computing, 1989
- Connectionism and cognitive architecture: A critical analysisCognition, 1988