VS3‐NET: Neural variational inference model for machine‐reading comprehension
We propose the VS3‐NET model to solve the task of question answering questions with machine‐reading comprehension that searches for an appropriate answer in a given context. VS3‐NET is a model that trains latent variables for each question using variational inferences based on a model of a simple re...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Electronics and Telecommunications Research Institute (ETRI)
2019-07-01
|
Series: | ETRI Journal |
Subjects: | |
Online Access: | https://doi.org/10.4218/etrij.2018-0467 |
id |
doaj-a28089dbf9e54571813db7b8e3470272 |
---|---|
record_format |
Article |
spelling |
doaj-a28089dbf9e54571813db7b8e34702722020-11-25T03:05:37ZengElectronics and Telecommunications Research Institute (ETRI)ETRI Journal1225-64631225-64632019-07-0141677178110.4218/etrij.2018-046710.4218/etrij.2018-0467VS3‐NET: Neural variational inference model for machine‐reading comprehensionCheoneum Park0Changki Lee1Heejun Song2Kangwon National UniversityKangwon National UniversitySamsung ResearchWe propose the VS3‐NET model to solve the task of question answering questions with machine‐reading comprehension that searches for an appropriate answer in a given context. VS3‐NET is a model that trains latent variables for each question using variational inferences based on a model of a simple recurrent unit‐based sentences and self‐matching networks. The types of questions vary, and the answers depend on the type of question. To perform efficient inference and learning, we introduce neural question‐type models to approximate the prior and posterior distributions of the latent variables, and we use these approximated distributions to optimize a reparameterized variational lower bound. The context given in machine‐reading comprehension usually comprises several sentences, leading to performance degradation caused by context length. Therefore, we model a hierarchical structure using sentence encoding, in which as the context becomes longer, the performance degrades. Experimental results show that the proposed VS3‐NET model has an exact‐match score of 76.8% and an F1 score of 84.5% on the SQuAD test set.https://doi.org/10.4218/etrij.2018-0467machine reading comprehensionquestion answeringsquadvariational inferencevs3-net |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Cheoneum Park Changki Lee Heejun Song |
spellingShingle |
Cheoneum Park Changki Lee Heejun Song VS3‐NET: Neural variational inference model for machine‐reading comprehension ETRI Journal machine reading comprehension question answering squad variational inference vs3-net |
author_facet |
Cheoneum Park Changki Lee Heejun Song |
author_sort |
Cheoneum Park |
title |
VS3‐NET: Neural variational inference model for machine‐reading comprehension |
title_short |
VS3‐NET: Neural variational inference model for machine‐reading comprehension |
title_full |
VS3‐NET: Neural variational inference model for machine‐reading comprehension |
title_fullStr |
VS3‐NET: Neural variational inference model for machine‐reading comprehension |
title_full_unstemmed |
VS3‐NET: Neural variational inference model for machine‐reading comprehension |
title_sort |
vs3‐net: neural variational inference model for machine‐reading comprehension |
publisher |
Electronics and Telecommunications Research Institute (ETRI) |
series |
ETRI Journal |
issn |
1225-6463 1225-6463 |
publishDate |
2019-07-01 |
description |
We propose the VS3‐NET model to solve the task of question answering questions with machine‐reading comprehension that searches for an appropriate answer in a given context. VS3‐NET is a model that trains latent variables for each question using variational inferences based on a model of a simple recurrent unit‐based sentences and self‐matching networks. The types of questions vary, and the answers depend on the type of question. To perform efficient inference and learning, we introduce neural question‐type models to approximate the prior and posterior distributions of the latent variables, and we use these approximated distributions to optimize a reparameterized variational lower bound. The context given in machine‐reading comprehension usually comprises several sentences, leading to performance degradation caused by context length. Therefore, we model a hierarchical structure using sentence encoding, in which as the context becomes longer, the performance degrades. Experimental results show that the proposed VS3‐NET model has an exact‐match score of 76.8% and an F1 score of 84.5% on the SQuAD test set. |
topic |
machine reading comprehension question answering squad variational inference vs3-net |
url |
https://doi.org/10.4218/etrij.2018-0467 |
work_keys_str_mv |
AT cheoneumpark vs3netneuralvariationalinferencemodelformachinereadingcomprehension AT changkilee vs3netneuralvariationalinferencemodelformachinereadingcomprehension AT heejunsong vs3netneuralvariationalinferencemodelformachinereadingcomprehension |
_version_ |
1724677456117366784 |