The World is Not Binary: Learning to Rank with Grayscale Data for Dialogue Response Selection

  title={The World is Not Binary: Learning to Rank with Grayscale Data for Dialogue Response Selection},
  author={Zibo Lin and Deng Cai and Yan Wang and Xiaojiang Liu and Haitao Zheng and Shuming Shi},
  • Zibo Lin, Deng Cai, +3 authors Shuming Shi
  • Published in EMNLP 2020
  • Computer Science
  • Response selection plays a vital role in building retrieval-based conversation systems. Despite that response selection is naturally a learning-to-rank problem, most prior works take a point-wise view and train binary classifiers for this task: each response candidate is labeled either relevant (one) or irrelevant (zero). On the one hand, this formalization can be sub-optimal due to its ignorance of the diversity of response quality. On the other hand, annotating grayscale data for learning-to… CONTINUE READING
    2 Citations

    Figures and Tables from this paper

    Dialogue Response Selection with Hierarchical Curriculum Learning
    • PDF
    Multi-turn Dialogue Reading Comprehension with Pivot Turns and Knowledge
    • PDF


    Are Training Samples Correlated? Learning to Generate Dialogue Responses with Multiple References
    • 20
    • PDF
    Constructing Interpretive Spatio-Temporal Features for Multi-Turn Responses Selection
    • 9
    • PDF
    Learning to Respond with Deep Neural Networks for Retrieval-Based Human-Computer Conversation System
    • 225
    • Highly Influential
    • PDF
    Skeleton-to-Response: Dialogue Generation Guided by Retrieval Memory
    • 25
    • PDF
    A Dataset for Research on Short-Text Conversations
    • 134
    • PDF
    One Time of Interaction May Not Be Enough: Go Deep with an Interaction-over-Interaction Network for Response Selection in Dialogues
    • 44
    • Highly Influential
    • PDF
    Multi-Turn Response Selection for Chatbots with Deep Attention Matching Network
    • 155
    • Highly Influential
    • PDF
    Multi-view Response Selection for Human-Computer Conversation
    • 140
    • Highly Influential
    • PDF
    LSTM-based Deep Learning Models for non-factoid answer selection
    • 336
    • PDF