Conference paper Open Access

A Case Study of Closed-Domain Response Suggestion with Limited Training Data

Galke, Lukas; Gerstenkorn, Gunnar; Scherp, Ansgar


Citation Style Language JSON Export

{
  "DOI": "10.1007/978-3-319-99133-7_18", 
  "language": "eng", 
  "title": "A Case Study of Closed-Domain Response Suggestion with Limited Training Data", 
  "issued": {
    "date-parts": [
      [
        2018, 
        9, 
        6
      ]
    ]
  }, 
  "abstract": "<p>We analyze the problem of response suggestion in a closed domain along a real-world scenario of a digital library. We present a text-processing pipeline to generate question-answer pairs from chat transcripts. On this limited amount of training data, we compare retrieval-based, conditioned-generation, and dedicated representation learning approaches for response suggestion. Our results show that retrieval-based methods that strive to find similar, known contexts are preferable over parametric approaches from the conditioned-generation family, when the training data is limited. We, however, identify a specific representation learning approach that is competitive to the retrieval-based approaches despite the training data limitation.</p>", 
  "author": [
    {
      "family": "Galke, Lukas"
    }, 
    {
      "family": "Gerstenkorn, Gunnar"
    }, 
    {
      "family": "Scherp, Ansgar"
    }
  ], 
  "note": "This is a post-peer-review, pre-copyedit version of a paper published in Elloumi M, Granitzer M, Hameurlain A, Seifert C, Stein\nB, Tjoa A & Wagner R (eds.) Database and Expert Systems Applications. DEXA 2018. Communications in Computer and\nInformation Science, 903. The final authenticated version is available online at: https://doi.org/10.1007/978-3-319-99133-7_18", 
  "type": "paper-conference", 
  "id": "2583130"
}
17
20
views
downloads
Views 17
Downloads 20
Data volume 4.7 MB
Unique views 15
Unique downloads 18

Share

Cite as