arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.01190cs.CL
收藏

LSTM语言模型中主谓数一致性的影响路径

Influence Paths for Characterizing Subject-Verb Number Agreement in LSTM Language Models

Carnegie Mellon University(卡内基梅隆大学)

Kaiji Lu, Piotr Mardziel, Klas Leino, Matt Fedrikson, Anupam Datta

摘要:

基于LSTM的递归神经网络是许多自然语言处理(NLP)任务的最新技术。尽管他们的表现,目前还不清楚是否,或如何,LSTM学习自然语言的结构特征,如英语中的主谓数一致。由于缺乏这种理解,LSTM在该任务上的性能的一般性及其对相关任务的适用性仍然不确定。此外,错误不能适当地归因于缺乏结构能力、训练数据遗漏或其他异常故障。我们介绍 * 影响路径 *,因果关系的结构属性进行的路径跨越门和神经元的循环神经网络。该方法将影响的概念(主语的语法数对后继动词的语法数有影响)细化为一组门或神经元级路径。该集合将概念局部化和分段(例如,主谓一致),其构成要素(例如,受试者),以及相关或干扰元素(例如,吸引子)。我们在一个广泛研究的多层LSTM语言模型上验证了该方法,证明了它对主语-动词数一致性的解释。与基于诊断分类器和消融的先前结果相比,结果提供了LSTM对英语语言的这种结构方面的处理的更精细和更完整的视图。

英文摘要:

LSTM-based recurrent neural networks are the state-of-the-art for many natural language processing (NLP) tasks. Despite their performance, it is unclear whether, or how, LSTMs learn structural features of natural languages such as subject-verb number agreement in English. Lacking this understanding, the generality of LSTM performance on this task and their suitability for related tasks remains uncertain. Further, errors cannot be properly attributed to a lack of structural capability, training data omissions, or other exceptional faults. We introduce *influence paths*, a causal account of structural properties as carried by paths across gates and neurons of a recurrent neural network. The approach refines the notion of influence (the subject's grammatical number has influence on the grammatical number of the subsequent verb) into a set of gate or neuron-level paths. The set localizes and segments the concept (e.g., subject-verb agreement), its constituent elements (e.g., the subject), and related or interfering elements (e.g., attractors). We exemplify the methodology on a widely-studied multi-layer LSTM language model, demonstrating its accounting for subject-verb number agreement. The results offer both a finer and a more complete view of an LSTM's handling of this structural aspect of the English language than prior results based on diagnostic classifiers and ablation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑