Conference paper Open Access

Lexical Micro-adaptation for Neural Machine Translation

Xu, Jitao; Crego, Josep; Senellart, Jean


JSON Export

{
  "files": [
    {
      "links": {
        "self": "https://zenodo.org/api/files/4a8547db-fdad-44ec-a6b5-6651ba11ae01/IWSLT2019_paper_9.pdf"
      }, 
      "checksum": "md5:b145694f9a8fd725c74837681d966e22", 
      "bucket": "4a8547db-fdad-44ec-a6b5-6651ba11ae01", 
      "key": "IWSLT2019_paper_9.pdf", 
      "type": "pdf", 
      "size": 432333
    }
  ], 
  "owners": [
    50447
  ], 
  "doi": "10.5281/zenodo.3524977", 
  "stats": {
    "version_unique_downloads": 80.0, 
    "unique_views": 102.0, 
    "views": 120.0, 
    "version_views": 120.0, 
    "unique_downloads": 80.0, 
    "version_unique_views": 102.0, 
    "volume": 39342303.0, 
    "version_downloads": 91.0, 
    "downloads": 91.0, 
    "version_volume": 39342303.0
  }, 
  "links": {
    "doi": "https://doi.org/10.5281/zenodo.3524977", 
    "conceptdoi": "https://doi.org/10.5281/zenodo.3524976", 
    "bucket": "https://zenodo.org/api/files/4a8547db-fdad-44ec-a6b5-6651ba11ae01", 
    "conceptbadge": "https://zenodo.org/badge/doi/10.5281/zenodo.3524976.svg", 
    "html": "https://zenodo.org/record/3524977", 
    "latest_html": "https://zenodo.org/record/3524977", 
    "badge": "https://zenodo.org/badge/doi/10.5281/zenodo.3524977.svg", 
    "latest": "https://zenodo.org/api/records/3524977"
  }, 
  "conceptdoi": "10.5281/zenodo.3524976", 
  "created": "2019-11-01T12:12:12.518944+00:00", 
  "updated": "2020-01-20T17:16:34.832476+00:00", 
  "conceptrecid": "3524976", 
  "revision": 3, 
  "id": 3524977, 
  "metadata": {
    "access_right_category": "success", 
    "doi": "10.5281/zenodo.3524977", 
    "description": "<p>This work is inspired by a typical machine translation industry scenario in which translators make use of in-domain data for facilitating translation of similar or repeating sentences. We introduce a generic framework applied at inference in which a subset of segment pairs are first extracted from training data according to their similarity to the input sentences. These segments are then used to dynamically update the parameters of a generic NMT network, thus performing a&nbsp;lexical micro-adaptation. Our approach demonstrates strong adaptation performance to new and existing datasets including pseudo in-domain data. We evaluate our approach on a heterogeneous English-French training dataset showing accuracy gains on all evaluated domains when compared to strong adaptation baselines.</p>", 
    "language": "eng", 
    "title": "Lexical Micro-adaptation for Neural Machine Translation", 
    "license": {
      "id": "CC-BY-4.0"
    }, 
    "relations": {
      "version": [
        {
          "count": 1, 
          "index": 0, 
          "parent": {
            "pid_type": "recid", 
            "pid_value": "3524976"
          }, 
          "is_last": true, 
          "last_child": {
            "pid_type": "recid", 
            "pid_value": "3524977"
          }
        }
      ]
    }, 
    "communities": [
      {
        "id": "iwslt2019"
      }
    ], 
    "publication_date": "2019-11-02", 
    "creators": [
      {
        "affiliation": "SYSTRAN, 5 rue Feydeau, 75002 Paris (France)", 
        "name": "Xu, Jitao"
      }, 
      {
        "affiliation": "SYSTRAN, 5 rue Feydeau, 75002 Paris (France)", 
        "name": "Crego, Josep"
      }, 
      {
        "affiliation": "SYSTRAN, 5 rue Feydeau, 75002 Paris (France)", 
        "name": "Senellart, Jean"
      }
    ], 
    "access_right": "open", 
    "resource_type": {
      "subtype": "conferencepaper", 
      "type": "publication", 
      "title": "Conference paper"
    }, 
    "related_identifiers": [
      {
        "scheme": "doi", 
        "identifier": "10.5281/zenodo.3524976", 
        "relation": "isVersionOf"
      }
    ]
  }
}
120
91
views
downloads
All versions This version
Views 120120
Downloads 9191
Data volume 39.3 MB39.3 MB
Unique views 102102
Unique downloads 8080

Share

Cite as