Multi-Perspective Context Matching for Machine Comprehension

By Zhiguo Wang, Haitao Mi, Wael Hamza, Radu Florian. 2016

158 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Retrieval & Search, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Retrieval & Search

In Natural Language Processing

Citations in the corpus

NewsQA: A Machine Comprehension Dataset (2016, methodology): “The SQuAD authors measured human accuracy at 0.905 in F1 (we measured human F1 at 0.807 using a different methodology); at the time of writing, the strongest published model to date achieves 0.778 F1 (Wang et al., 2016).”
Coarse-to-Fine Question Answering for Long Documents (2016, background): “, 2015) is a synthetic dataset that captures various aspects of reasoning; and SQuAD (Rajpurkar et al., 2016; Wang et al., 2016; Xiong et al., 2016) and NewsQA (Trischler et al.”
Multi-passage BERT: A Globally Normalized BERT Model for Open-domain Question Answering (2019, background): “Fourth , most effective QA and RC models highly rely on explicit inter-sentence matching between questions and passages (Wang and Jiang, 2017; Wang et al., 2016; Seo et al., 2017; Wang et al., 2017), whereas BERT only applies self-attention layers over the concatenation of a question-passage pair.”
Zero-Shot Relation Extraction via Reading Comprehension (2017, background): “The task of obtaining answer spans to natural-language questions has been recently studied on the SQuAD dataset (Rajpurkar et al., 2016; Xiong et al., 2016; Lee et al., 2016; Wang et al., 2016).”
Semi-Supervised QA with Generative Domain-Adaptive Nets (2017, background): “Various neural models based on attention mechanisms (Wang and Jiang, 2016; Seo et al., 2016; Xiong et al., 2016; Wang et al., 2016; Dhingra et al., 2016; Kadlec et al., 2016; Trischler et al., 2016b; Sordoni et al., 2016; Cui et al., 2016; Chen et al., 2016) have been proposed to tackle the tasks of”
Adversarial Examples for Evaluating Reading Comprehension Systems (2017, background): “, 2017), Ruminating Reader (Gong and Bowman, 2017), MultiPerspective Context Matching (MPCM) Single version (Wang et al., 2016), RaSOR (Lee et al.”
Gated Self-Matching Networks for Reading Comprehension and Question Answering (2017, background): “…Dynamic Coattention Networks (Xiong et al., 2016) 65.4 / 75.6 66.2 / 75.9 RaSoR (Lee et al., 2016) 66.4 / 74.9 - / - BiDAF (Seo et al., 2016) 68.0 / 77.3 68.0 / 77.3 jNet (Zhang et al., 2017) - / - 68.7 / 77.4 Multi-Perspective Matching (Wang et al., 2016) - / - 68.9 / 77.8 FastQA (Weissenborn et”
Making Neural QA as Simple as Possible but not Simpler (2017, background): “Feeding additional word-level features similar to ours is rarely done with the exception of Wang et al. (2017); Li et al. (2016).”
FusionNet: Fusing via Fully-Aware Attention with Application to Machine Comprehension (2017, background): “, 2017a) D MPCM (Wang et al., 2016) D D Mnemonic Reader (Hu et al.”
Exploring Question Understanding and Adaptation in Neural-Network-Based Question Answering (2017, neutral): “Logistic Regression Baseline [Rajpurkar et al., 2016] 40.4 51.0 Match-LSTM with Ans-Ptr (Sentence) [Wang and Jiang, 2016] 54.505 67.748 Match-LSTM with Ans-Ptr (Boundary) [Wang and Jiang, 2016] 60.474 70.695 Dynamic Chunk Reader [Yu et al., 2016] 62.499 70.956 Fine-Grained Gating [Yang et al., 2016]”
R3: Reinforced Reader-Ranker for Open-Domain Question Answering (2017, methodology): “Dividing the pipeline into IR and RC stages leverages an enormous body of research in both IR and RC, including recent successes in RC via neural network techniques (Wang and Jiang, 2017b; Yu et al., 2016; Wang et al., 2016; Xiong et al., 2017; Wang et al., 2017).”
MEMEN: Multi-layer Embedding with Memory Networks for Machine Comprehension (2017, background): “Many works are based on the task of machine reading comprehension, and attention mechanism have been particularly successful(Xiong et al., 2016; Cui et al., 2016; Wang et al., 2016; Seo et al., 2016; Hu et al., 2017; Shen et al., 2016; Wang et al., 2017).”
Knowledgeable Reader: Enhancing Cloze-Style Reading Comprehension with External Commonsense Knowledge (2018, methodology): “…hops and modeling a deeper relation between question and document was further developed by several models (Seo et al., 2017; Xiong et al., 2016; Wang et al., 2016, 2017; Shen et al., 2016) on another generation of RC datasets, e.g. SQuAD (Rajpurkar et al., 2016), NewsQA (Trischler et al.,…”
Reinforced Mnemonic Reader for Machine Comprehension (2017, methodology): “…and Courville 2016) and large-scale MC benchmark datasets (Hermann et al. 2015; Hill et al. 2016; Rajpurkar et al. 2016; Joshi et al. 2017), end-to-end neural networks have achieved promising results on MC tasks (Wang et al. 2016; Xiong, Zhong, and Socher 2017; Seo et al. 2017; Wang et al. 2017).”
Contextualized Word Representations for Reading Comprehension (2017, methodology): “ + TR [3] 44.5 53.9 MPCM [4] 40.3 50.0 RaSoR (base model) [5] 39.5 49.5 ReasoNet [6] 39.4 50.3 jNet [7] 37.9 47.0 Table 3: Single-model F1 on adversarial SQuAD. [1,3] This work. [2]Hu et al.(2017) [4]Wang et al. (2016) [5]Lee et al.(2016) [6]Shen et al.(2017) [7] Zhang et al.(2017) 5 Experimental se”