A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media
Along with the development of social media on the internet, dialogue systems are becoming more and more intelligent to meet users’ needs for communication, emotion, and social intercourse. Previous studies usually use sequence-to-sequence learning with recurrent neural networks for response generati...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2020-01-01
|
| Series: | Complexity |
| Online Access: | http://dx.doi.org/10.1155/2020/3710104 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832554067911507968 |
|---|---|
| author | Yuanyuan Cai Min Zuo Qingchuan Zhang Haitao Xiong Ke Li |
| author_facet | Yuanyuan Cai Min Zuo Qingchuan Zhang Haitao Xiong Ke Li |
| author_sort | Yuanyuan Cai |
| collection | DOAJ |
| description | Along with the development of social media on the internet, dialogue systems are becoming more and more intelligent to meet users’ needs for communication, emotion, and social intercourse. Previous studies usually use sequence-to-sequence learning with recurrent neural networks for response generation. However, recurrent-based learning models heavily suffer from the problem of long-distance dependencies in sequences. Moreover, some models neglect crucial information in the dialogue contexts, which leads to uninformative and inflexible responses. To address these issues, we present a bichannel transformer with context encoding (BCTCE) for document-driven conversation. This conversational generator consists of a context encoder, an utterance encoder, and a decoder with attention mechanism. The encoders aim to learn the distributed representation of input texts. The multihop attention mechanism is used in BCTCE to capture the interaction between documents and dialogues. We evaluate the proposed BCTCE by both automatic evaluation and human judgment. The experimental results on the dataset CMU_DoG indicate that the proposed model yields significant improvements over the state-of-the-art baselines on most of the evaluation metrics, and the generated responses of BCTCE are more informative and more relevant to dialogues than baselines. |
| format | Article |
| id | doaj-art-20cf92edb91e4b3888bd47f0d7a58701 |
| institution | Kabale University |
| issn | 1076-2787 1099-0526 |
| language | English |
| publishDate | 2020-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Complexity |
| spelling | doaj-art-20cf92edb91e4b3888bd47f0d7a587012025-02-03T05:52:29ZengWileyComplexity1076-27871099-05262020-01-01202010.1155/2020/37101043710104A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social MediaYuanyuan Cai0Min Zuo1Qingchuan Zhang2Haitao Xiong3Ke Li4National Engineering Laboratory for Agri-Product Quality Traceability and Beijing Key Laboratory of Big Data Technology for Food Safety, Beijing Technology and Business University, Beijing 100048, ChinaNational Engineering Laboratory for Agri-Product Quality Traceability and Beijing Key Laboratory of Big Data Technology for Food Safety, Beijing Technology and Business University, Beijing 100048, ChinaNational Engineering Laboratory for Agri-Product Quality Traceability and Beijing Key Laboratory of Big Data Technology for Food Safety, Beijing Technology and Business University, Beijing 100048, ChinaNational Engineering Laboratory for Agri-Product Quality Traceability and Beijing Key Laboratory of Big Data Technology for Food Safety, Beijing Technology and Business University, Beijing 100048, ChinaBeijing University of Posts and Telecommunications, No. 10 Xitucheng Road, Haidian District, Beijing, ChinaAlong with the development of social media on the internet, dialogue systems are becoming more and more intelligent to meet users’ needs for communication, emotion, and social intercourse. Previous studies usually use sequence-to-sequence learning with recurrent neural networks for response generation. However, recurrent-based learning models heavily suffer from the problem of long-distance dependencies in sequences. Moreover, some models neglect crucial information in the dialogue contexts, which leads to uninformative and inflexible responses. To address these issues, we present a bichannel transformer with context encoding (BCTCE) for document-driven conversation. This conversational generator consists of a context encoder, an utterance encoder, and a decoder with attention mechanism. The encoders aim to learn the distributed representation of input texts. The multihop attention mechanism is used in BCTCE to capture the interaction between documents and dialogues. We evaluate the proposed BCTCE by both automatic evaluation and human judgment. The experimental results on the dataset CMU_DoG indicate that the proposed model yields significant improvements over the state-of-the-art baselines on most of the evaluation metrics, and the generated responses of BCTCE are more informative and more relevant to dialogues than baselines.http://dx.doi.org/10.1155/2020/3710104 |
| spellingShingle | Yuanyuan Cai Min Zuo Qingchuan Zhang Haitao Xiong Ke Li A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media Complexity |
| title | A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media |
| title_full | A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media |
| title_fullStr | A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media |
| title_full_unstemmed | A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media |
| title_short | A Bichannel Transformer with Context Encoding for Document-Driven Conversation Generation in Social Media |
| title_sort | bichannel transformer with context encoding for document driven conversation generation in social media |
| url | http://dx.doi.org/10.1155/2020/3710104 |
| work_keys_str_mv | AT yuanyuancai abichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT minzuo abichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT qingchuanzhang abichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT haitaoxiong abichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT keli abichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT yuanyuancai bichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT minzuo bichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT qingchuanzhang bichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT haitaoxiong bichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia AT keli bichanneltransformerwithcontextencodingfordocumentdrivenconversationgenerationinsocialmedia |