{ "access": { "embargo": { "active": false, "reason": null }, "files": "public", "record": "public", "status": "open" }, "created": "2019-11-01T13:23:33.239846+00:00", "custom_fields": {}, "deletion_status": { "is_deleted": false, "status": "P" }, "files": { "count": 1, "enabled": true, "entries": { "IWSLT2019_paper_23.pdf": { "checksum": "md5:8aa1a0901fb5a22f51a9126b88e71ae0", "ext": "pdf", "id": "7ee4e25d-a777-4605-bb00-030fc464a518", "key": "IWSLT2019_paper_23.pdf", "metadata": null, "mimetype": "application/pdf", "size": 311102 } }, "order": [], "total_bytes": 311102 }, "id": "3525026", "is_draft": false, "is_published": true, "links": { "access": "https://zenodo.org/api/records/3525026/access", "access_links": "https://zenodo.org/api/records/3525026/access/links", "access_request": "https://zenodo.org/api/records/3525026/access/request", "access_users": "https://zenodo.org/api/records/3525026/access/users", "archive": "https://zenodo.org/api/records/3525026/files-archive", "archive_media": "https://zenodo.org/api/records/3525026/media-files-archive", "communities": "https://zenodo.org/api/records/3525026/communities", "communities-suggestions": "https://zenodo.org/api/records/3525026/communities-suggestions", "doi": "https://doi.org/10.5281/zenodo.3525026", "draft": "https://zenodo.org/api/records/3525026/draft", "files": "https://zenodo.org/api/records/3525026/files", "latest": "https://zenodo.org/api/records/3525026/versions/latest", "latest_html": "https://zenodo.org/records/3525026/latest", "media_files": "https://zenodo.org/api/records/3525026/media-files", "parent": "https://zenodo.org/api/records/3525025", "parent_doi": "https://zenodo.org/doi/10.5281/zenodo.3525025", "parent_html": "https://zenodo.org/records/3525025", "requests": "https://zenodo.org/api/records/3525026/requests", "reserve_doi": "https://zenodo.org/api/records/3525026/draft/pids/doi", "self": "https://zenodo.org/api/records/3525026", "self_doi": "https://zenodo.org/doi/10.5281/zenodo.3525026", "self_html": "https://zenodo.org/records/3525026", "self_iiif_manifest": "https://zenodo.org/api/iiif/record:3525026/manifest", "self_iiif_sequence": "https://zenodo.org/api/iiif/record:3525026/sequence/default", "versions": "https://zenodo.org/api/records/3525026/versions" }, "media_files": { "count": 0, "enabled": false, "entries": {}, "order": [], "total_bytes": 0 }, "metadata": { "creators": [ { "affiliations": [ { "name": "LIMSI, CNRS, Universit\u00e9 Paris-Saclay, Ba\u02c6t. 508, rue John von Neumann, Campus Universitaire, F-91405 Orsay" } ], "person_or_org": { "family_name": "Ho", "given_name": "Anh Khoa Ngo", "name": "Ho, Anh Khoa Ngo", "type": "personal" } }, { "affiliations": [ { "name": "LIMSI, CNRS, Universit\u00e9 Paris-Saclay, Ba\u02c6t. 508, rue John von Neumann, Campus Universitaire, F-91405 Orsay" } ], "person_or_org": { "family_name": "Yvon", "given_name": "Fran\u00e7ois", "name": "Yvon, Fran\u00e7ois", "type": "personal" } } ], "description": "
Word alignments identify translational correspondences between words in a parallel sentence pair and is used, for instance, to learn bilingual dictionaries, to train statistical machine translation systems, or to perform quality estimation. In most areas of natural lan- guage processing, neural network models nowadays constitute the preferred approach, a situation that might also apply to word align- ment models. In this work, we study and comprehensively evaluate neural models for unsupervised word alignment for four language pairs, contrasting several variants of neural models. We show that in most settings, neural versions of the IBM-1 and hidden Markov models vastly outperform their discrete counterparts. We also analyze typical alignment errors of the baselines that our models over- come to illustrate the benefits — and the limitations — of these new models for morphologically rich languages.
", "publication_date": "2019-11-02", "publisher": "Zenodo", "resource_type": { "id": "publication-conferencepaper", "title": { "de": "Konferenzbeitrag", "en": "Conference paper" } }, "rights": [ { "description": { "en": "The Creative Commons Attribution license allows re-distribution and re-use of a licensed work on the condition that the creator is appropriately credited." }, "icon": "cc-by-icon", "id": "cc-by-4.0", "props": { "scheme": "spdx", "url": "https://creativecommons.org/licenses/by/4.0/legalcode" }, "title": { "en": "Creative Commons Attribution 4.0 International" } } ], "title": "Neural Baselines for Word Alignment" }, "parent": { "access": { "owned_by": { "user": 50447 } }, "communities": { "default": "db2a5261-ee99-4982-9879-e8274ea4f1af", "entries": [ { "access": { "member_policy": "open", "members_visibility": "public", "record_policy": "open", "review_policy": "open", "visibility": "public" }, "children": { "allow": false }, "created": "2019-11-01T07:43:01.061176+00:00", "custom_fields": {}, "deletion_status": { "is_deleted": false, "status": "P" }, "id": "db2a5261-ee99-4982-9879-e8274ea4f1af", "links": {}, "metadata": { "curation_policy": "Papers published at 16th International Workshop on Spoken Language Translation 2019 in Hong Kong, as maintained by the program and evaluation chairs.
\r\n", "description": "", "page": "", "title": "16th International Workshop on Spoken Language Translation 2019" }, "revision_id": 0, "slug": "iwslt2019", "updated": "2019-11-01T07:43:01.127561+00:00" } ], "ids": [ "db2a5261-ee99-4982-9879-e8274ea4f1af" ] }, "id": "3525025", "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.3525025", "provider": "datacite" } } }, "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.3525026", "provider": "datacite" }, "oai": { "identifier": "oai:zenodo.org:3525026", "provider": "oai" } }, "revision_id": 2, "stats": { "all_versions": { "data_volume": 36710036.0, "downloads": 118, "unique_downloads": 112, "unique_views": 184, "views": 192 }, "this_version": { "data_volume": 36710036.0, "downloads": 118, "unique_downloads": 112, "unique_views": 183, "views": 191 } }, "status": "published", "updated": "2020-01-20T17:28:58.998132+00:00", "versions": { "index": 1, "is_latest": true } }