Parallelograms revisited: Exploring the limitations of vector space models for simple analogies
Parallelograms revisited: Exploring the limitations of vector space models for simple analogies
复制标题
DOI：
10.1016/j.cognition.2020.104440
复制
发表时间：
2020-08
期刊：
影响因子：
3.4
通讯作者：
Joshua C. Peterson;Dawn Chen;T. Griffiths
中科院分区：
文献类型：
--
作者：
Joshua C. Peterson;Dawn Chen;T. Griffiths
经典心理学理论已经证明了空间表征的力量和局限性，为推理物体之间的相似性提供了几何工具，并表明人类的直觉有时会违反几何空间的约束。最近用于推导单词向量空间嵌入的机器学习方法已经开始引起人们的注意，因为它们惊人的能力在大型语料库中一致地捕捉简单的类比，赋予了由心理学家首次提出并简要探索的经典类比模型以平行四边形的新生命。我们评估了应用于现代数据驱动的词嵌入的平行四边形类比模型，详细分析了该方法在词对领域捕获人类行为的程度。使用一个大型的人类类比完成的基准数据集，我们表明，仅单词相似性就比平行四边形模型更好地捕捉到人类反应的某些方面。为了更好地了解这些模型对关系相似性的预测能力，我们还收集了大量人类关系相似性判断的数据集，发现平行四边形模型比其他模型更好地捕捉了一些语义关系。最后，我们证明了基于向量空间内在几何约束的平行四边形类比模型的更深层次的局限性，与经典的项目相似性结果相一致。综上所述，这些结果表明，虽然现代单词嵌入在规模上捕捉语义相似性方面做了令人印象深刻的工作，但仅靠平行四边形模型不足以解释人们如何形成即使是最简单的类比。
Classic psychological theories have demonstrated the power and limitations of spatial representations, providing geometric tools for reasoning about the similarity of objects and showing that human intuitions sometimes violate the constraints of geometric spaces. Recent machine learning methods for deriving vector-space embeddings of words have begun to garner attention for their surprising capacity to capture simple analogies consistently across large corpora, giving new life to a classic model of analogies as parallelograms that was first proposed and briefly explored by psychologists. We evaluate the parallelogram model of analogy as applied to modern data-driven word embeddings, providing a detailed analysis of the extent to which this approach captures human behavior in the domain of word pairs. Using a large novel benchmark dataset of human analogy completions, we show that word similarity alone surprisingly captures some aspects of human responses better than the parallelogram model. To gain a fine-grained picture of how well these models predict relational similarity, we also collect a large dataset of human relational similarity judgments and find that the parallelogram model captures some semantic relationships better than others. Finally, we provide evidence for deeper limitations of the parallelogram model of analogy based on the intrinsic geometric constraints of vector spaces, paralleling classic results for item similarity. Taken together, these results show that while modern word embeddings do an impressive job of capturing semantic similarity at scale, the parallelogram model alone is insufficient to account for how people form even the simplest analogies.