Measuring compositional generalization: A comprehensive method on realistic data
arXiv preprint arXiv:1912.09713, 2019•arxiv.org
State-of-the-art machine learning methods exhibit limited compositional generalization. At
the same time, there is a lack of realistic benchmarks that comprehensively measure this
ability, which makes it challenging to find and evaluate improvements. We introduce a novel
method to systematically construct such benchmarks by maximizing compound divergence
while guaranteeing a small atom divergence between train and test sets, and we
quantitatively compare this method to other approaches for creating compositional …
the same time, there is a lack of realistic benchmarks that comprehensively measure this
ability, which makes it challenging to find and evaluate improvements. We introduce a novel
method to systematically construct such benchmarks by maximizing compound divergence
while guaranteeing a small atom divergence between train and test sets, and we
quantitatively compare this method to other approaches for creating compositional …
State-of-the-art machine learning methods exhibit limited compositional generalization. At the same time, there is a lack of realistic benchmarks that comprehensively measure this ability, which makes it challenging to find and evaluate improvements. We introduce a novel method to systematically construct such benchmarks by maximizing compound divergence while guaranteeing a small atom divergence between train and test sets, and we quantitatively compare this method to other approaches for creating compositional generalization benchmarks. We present a large and realistic natural language question answering dataset that is constructed according to this method, and we use it to analyze the compositional generalization ability of three machine learning architectures. We find that they fail to generalize compositionally and that there is a surprisingly strong negative correlation between compound divergence and accuracy. We also demonstrate how our method can be used to create new compositionality benchmarks on top of the existing SCAN dataset, which confirms these findings.
arxiv.org
以上显示的是最相近的搜索结果。 查看全部搜索结果