A Conversational Agent Powered by Online Learning

被引:0
作者
Mendonca, Vania [1 ]
Melo, Francisco S. [1 ]
Coheur, Luisa [1 ]
Sardinha, Alberto [1 ]
机构
[1] INESC ID, Inst Super Tecn, Ave Prof Doutor Anibal Cavaco Silva, Porto Salvo, Portugal
来源
AAMAS'17: PROCEEDINGS OF THE 16TH INTERNATIONAL CONFERENCE ON AUTONOMOUS AGENTS AND MULTIAGENT SYSTEMS | 2017年
关键词
Conversational agents; online learning; Exponentially Weighted Average Forecaster;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
In this work, we improve the performance of a dialogue engine, Say Something Smart, using online learning. Given a request by a user, this engine selects an answer from a corpus of movie subtitles, weighting the quality of each candidate answer according to several criteria and selecting the one that is chosen by the most representative criteria. We contribute with an online approach, using sequential learning, that adjusts the weights of the different criteria using a reference corpus of actual dialogues as input to simulate user feedback. This approach effectively allowed Say Something Smart to improve its performance at each interaction, as shown in an experiment performed in a test corpus.
引用
收藏
页码:1637 / 1639
页数:3
相关论文
共 12 条
  • [1] Ameixa D, 2014, LECT NOTES ARTIF INT, V8637, P13, DOI 10.1007/978-3-319-09767-1_2
  • [2] [Anonymous], 2011, ACL
  • [3] Cesa-Bianchi N., 2006, PREDICTION LEARNING
  • [4] Cuayahuitl H., 2012, NAACL HLT WORKSH FUT, P7
  • [5] Fialho P., 2013, ACL 2013, P61
  • [6] Levin E, 2000, IEEE T SPEECH AUDIO, V8
  • [7] THE WEIGHTED MAJORITY ALGORITHM
    LITTLESTONE, N
    WARMUTH, MK
    [J]. INFORMATION AND COMPUTATION, 1994, 108 (02) : 212 - 261
  • [8] Magarreiro D., 2014, SEMDIAL 2014 DIALWAT
  • [9] Pietquin Olivier., 2011, IJCAI Proceedings-International Joint Conference on Artificial Intelligence, V22, page, P1878
  • [10] Optimizing dialogue management with reinforcement learning: Experiments with the NJFun system
    Singh, S
    Litman, D
    Kearns, M
    Walker, M
    [J]. JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 2002, 16 : 105 - 133