论文信息 - A Reinforcement Learning Approach for Dynamic Search

A Reinforcement Learning Approach for Dynamic Search

TREC Dynamic Domain (DD) track is intended to support the research in dynamic, exploratory search within complex domains. It simulates an interactive search process where the search system is expected to improve its efficiency and effectiveness based on its interaction with the user. We propose to model the dynamic search as a reinforcement learning problem and use neural network to find the best policy during a search process. We show a great potential of deep reinforcement learning on DD track.

Grace Hui Yang | Zhiwen Tang | G. Yang | Zhiwen Tang

[1] Grace Hui Yang,et al. Investigating per Topic Upper Bound for Session Search Evaluation , 2017, ICTIR.

[2] Yuxi Li,et al. Deep Reinforcement Learning: An Overview , 2017, ArXiv.

[3] 一樹美添,et al. 5分で分かる! ? 有名論文ナナメ読み：Silver, D. et al. : Mastering the Game of Go without Human Knowledge , 2018 .

[4] Yiming Yang,et al. Modeling Expected Utility of Multi-session Information Distillation , 2009, ICTIR.

[5] 悠太菊池,et al. 大規模要約資源としてのNew York Times Annotated Corpus , 2015 .

[6] Grace Hui Yang,et al. A Contextual Bandit Approach to Dynamic Search , 2017, ICTIR.

[7] Richard S. Sutton,et al. Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[8] Peter Dayan,et al. Q-learning , 1992, Machine Learning.

[9] Shane Legg,et al. Human-level control through deep reinforcement learning , 2015, Nature.

[10] Jürgen Schmidhuber,et al. Long Short-Term Memory , 1997, Neural Computation.

[11] Grace Hui Yang,et al. The water filling model and the cube test: multi-dimensional evaluation for professional search , 2013, CIKM.

[12] Jeffrey Dean,et al. Distributed Representations of Words and Phrases and their Compositionality , 2013, NIPS.

[13] Demis Hassabis,et al. Mastering the game of Go without human knowledge , 2017, Nature.

[14] Lois M. L. Delcambre,et al. Discounted Cumulated Gain Based Evaluation of Multiple-Query IR Sessions , 2008, ECIR.