Teaching machines to read and comprehend
Teaching machines to read natural language documents remains an elusive challenge. Machine reading systems can be tested on their ability to answer questions posed on the contents of documents that they have seen, but until now large scale training and test datasets have been missing for this type o...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Journal article |
| Published: |
Neural Information Processing Systems
2015
|
| _version_ | 1797051286479175680 |
|---|---|
| author | Hermann, K Kočiský, T Grefenstette, E Espeholt, L Kay, W Suleyman, M Blunsom, P |
| author_facet | Hermann, K Kočiský, T Grefenstette, E Espeholt, L Kay, W Suleyman, M Blunsom, P |
| author_sort | Hermann, K |
| collection | OXFORD |
| description | Teaching machines to read natural language documents remains an elusive challenge. Machine reading systems can be tested on their ability to answer questions posed on the contents of documents that they have seen, but until now large scale training and test datasets have been missing for this type of evaluation. In this work we define a new methodology that resolves this bottleneck and provides large scale supervised reading comprehension data. This allows us to develop a class of attention based deep neural networks that learn to read real documents and answer complex questions with minimal prior knowledge of language structure. |
| first_indexed | 2024-03-06T18:17:19Z |
| format | Journal article |
| id | oxford-uuid:050e7840-1ff3-49db-8d36-e83ed0adf8f7 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T18:17:19Z |
| publishDate | 2015 |
| publisher | Neural Information Processing Systems |
| record_format | dspace |
| spelling | oxford-uuid:050e7840-1ff3-49db-8d36-e83ed0adf8f72022-03-26T08:55:03ZTeaching machines to read and comprehendJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:050e7840-1ff3-49db-8d36-e83ed0adf8f7Symplectic Elements at OxfordNeural Information Processing Systems2015Hermann, KKočiský, TGrefenstette, EEspeholt, LKay, WSuleyman, MBlunsom, PTeaching machines to read natural language documents remains an elusive challenge. Machine reading systems can be tested on their ability to answer questions posed on the contents of documents that they have seen, but until now large scale training and test datasets have been missing for this type of evaluation. In this work we define a new methodology that resolves this bottleneck and provides large scale supervised reading comprehension data. This allows us to develop a class of attention based deep neural networks that learn to read real documents and answer complex questions with minimal prior knowledge of language structure. |
| spellingShingle | Hermann, K Kočiský, T Grefenstette, E Espeholt, L Kay, W Suleyman, M Blunsom, P Teaching machines to read and comprehend |
| title | Teaching machines to read and comprehend |
| title_full | Teaching machines to read and comprehend |
| title_fullStr | Teaching machines to read and comprehend |
| title_full_unstemmed | Teaching machines to read and comprehend |
| title_short | Teaching machines to read and comprehend |
| title_sort | teaching machines to read and comprehend |
| work_keys_str_mv | AT hermannk teachingmachinestoreadandcomprehend AT kociskyt teachingmachinestoreadandcomprehend AT grefenstettee teachingmachinestoreadandcomprehend AT espeholtl teachingmachinestoreadandcomprehend AT kayw teachingmachinestoreadandcomprehend AT suleymanm teachingmachinestoreadandcomprehend AT blunsomp teachingmachinestoreadandcomprehend |