Making Transformers Solve Compositional Tasks
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)2022
Citations Over TimeTop 10% of 2022 papers
Abstract
Several studies have reported the inability of Transformer models to generalize compositionally, a key type of generalization in many NLP tasks such as semantic parsing. In this paper we explore the design space of Transformer models showing that the inductive biases given to the model by several design decisions significantly impact compositional generalization. We identified Transformer configurations that generalize compositionally significantly better than previously reported in the literature in many compositional tasks. We achieve state-of-the-art results in a semantic parsing compositional generalization benchmark (COGS), and a string edit operation composition benchmark (PCFG).
Related Papers
- → A Benchmark Test Structure for Experimental Dynamic Substructuring(2011)9 cited
- → Solutions to the Third Benchmark Control Problem(1991)3 cited
- Theoretical Analysis of the Benchmark for Choosing Manipulative Instruments of Monetary Policies(2009)
- → Exploring disk performance benchmarks(2017)
- → Support Structure Performance Benchmark(2023)