DC Field | Value | Language |
---|---|---|
dc.contributor.author | Seonwoo, Yeon | ko |
dc.contributor.author | Lee, Sang-Woo | ko |
dc.contributor.author | Oh, Alice Haeyun | ko |
dc.contributor.author | Ha, Jung-Woo | ko |
dc.contributor.author | Kim, Ji-Hoon | ko |
dc.date.accessioned | 2021-11-10T06:49:54Z | - |
dc.date.available | 2021-11-10T06:49:54Z | - |
dc.date.created | 2021-11-02 | - |
dc.date.issued | 2021-08 | - |
dc.identifier.citation | The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021) | - |
dc.identifier.uri | http://hdl.handle.net/10203/289122 | - |
dc.description.abstract | In multi-hop QA, answering complex questions entails iterative document retrieval for finding the missing entity of the question. The main steps of this process are sub-question detection, document retrieval for the subquestion, and generation of a new query for the final document retrieval. However, building a dataset that contains complex questions with sub-questions and their corresponding documents requires costly human annotation. To address the issue, we propose a new method for weakly supervised multi-hop retriever pretraining without human efforts. Our method includes 1) a pre-training task for generating vector representations of complex questions, 2) a scalable data generation method that produces the nested structure of question and subquestion as weak supervision for pre-training, and 3) a pre-training model structure based on dense encoders. We conduct experiments to compare the performance of our pre-trained retriever with several state-of-the-art models on end-to-end multi-hop QA as well as document retrieval. The experimental results show that our pre-trained retriever is effective and also robust on limited data and computational resources. | - |
dc.language | English | - |
dc.publisher | Association for Computational Linguistics (ACL 2021) | - |
dc.title | Weakly Supervised Pre-Training for Multi-Hop Retriever | - |
dc.type | Conference | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021) | - |
dc.identifier.conferencecountry | TH | - |
dc.identifier.conferencelocation | The Berkeley Hotel, Bangkok | - |
dc.contributor.localauthor | Oh, Alice Haeyun | - |
dc.contributor.nonIdAuthor | Lee, Sang-Woo | - |
dc.contributor.nonIdAuthor | Ha, Jung-Woo | - |
dc.contributor.nonIdAuthor | Kim, Ji-Hoon | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.