arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1612.04211cs.CL
收藏

面向机器理解的多视角上下文匹配

Multi-Perspective Context Matching for Machine Comprehension

IBM T.J. Watson Research Center(IBM T.J. 沃森研究中心)

Zhiguo Wang, Haitao Mi, Wael Hamza, Radu Florian

摘要:

以前的机器理解(MC)数据集要么太小,无法训练端到端深度学习模型,要么不够困难,无法评估当前MC技术的能力。新发布的SQuAD数据集克服了这些限制,并使我们有机会开发更真实的MC模型。基于此数据集,我们提出了一个多视角上下文匹配(MPCM)模型,这是一个端到端的系统,直接预测答案的开始和结束点在一个通道。我们的模型首先通过乘以针对问题计算的相关性权重来调整段落中的每个词嵌入向量。然后,我们使用双向LSTM对问题和加权段落进行编码。对于文章中的每个点,我们的模型从多个角度将该点的上下文与编码问题进行匹配,并生成匹配向量。给定这些匹配的向量,我们使用另一个双向LSTM来聚合所有信息并预测起点和终点。在SQuAD测试集上的实验结果表明,该模型在排行榜上取得了较好的效果。

英文摘要:

Previous machine comprehension (MC) datasets are either too small to train end-to-end deep learning models, or not difficult enough to evaluate the ability of current MC techniques. The newly released SQuAD dataset alleviates these limitations, and gives us a chance to develop more realistic MC models. Based on this dataset, we propose a Multi-Perspective Context Matching (MPCM) model, which is an end-to-end system that directly predicts the answer beginning and ending points in a passage. Our model first adjusts each word-embedding vector in the passage by multiplying a relevancy weight computed against the question. Then, we encode the question and weighted passage by using bi-directional LSTMs. For each point in the passage, our model matches the context of this point against the encoded question from multiple perspectives and produces a matching vector. Given those matched vectors, we employ another bi-directional LSTM to aggregate all the information and predict the beginning and ending points. Experimental result on the test set of SQuAD shows that our model achieves a competitive result on the leaderboard.

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。

↑