A Deep Architecture for Semantic Matching with Multiple Positional Sentence Representations
Shengxian Wan∗, Yanyan Lan†, Jiafeng Guo†, Jun Xu†, Liang Pang∗ Xueqi Cheng† CAS Key Lab of Network Data Science and Technology Institute of Computing Technology, Chinese Academy of Sciences, China ∗{wanshengxian, †{lanyanyan, guojiafeng, junxu,
Abstract
Matching natural language sentences is central for many applications such as information retrieval and question answering. Existing deep models rely on a single sentence representation or multiple granularity representations for matching. However, such methods cannot well capture the contextualized local information in the matching process. To tackle this problem, we present a new deep architecture to match two sentences with multiple positional sentence representations. Specifically, each positional sentence representation is a sentence representation at this position, generated by a bidirectional long short term memory (Bi-LSTM). The matching score is finally produced by aggregating interactions between these different positional sentence representations, through $k$ -Max pooling and a multi-layer perceptron. Our model has several advantages: (1) By using Bi-LSTM, rich context of the whole sentence is leveraged to capture the contextualized local information in each positional sentence representation; (2) By matching with multiple positional sentence representations, it is flexible to aggregate different important contextualized local information in a sentence to support the matc
中文速览
句子语义匹配（semantic matching）是问答系统、信息检索等任务的核心难题，现有方法要么把整句压缩成单一向量而丢失局部信息，要么只用词、短语等固定粒度的表示而无法感知完整句子的上下文语义。为此，研究者提出了 MV-LSTM 模型，其核心思路是用双向长短期记忆网络（Bi-LSTM）在句子的每个位置生成一个"位置句子表示"（positional sentence representation），每个表示都融合了来自全句两个方向的上下文信息，从而让同一个关键词在不同语境下能被区分对待。模型随后通过余弦、双线性或张量层等函数计算两句所有位置表示之间的交互矩阵，再用 k-Max 池化自动筛选最强的 k 个匹配信号，最终经多层感知机输出匹配分数。在问答和句子补全两类任务上的实验表明，MV-LSTM 全面超越了 ARC-I/II、CNTN、DeepMatch 等多个基线模型，证明了利用多位置上下文化句子表示进行匹配这一思路的有效性。
原文 arXiv:1511.08277;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1511.08277v1