Improving the Performance of the Reinforcement Learning Model for Answering Complex Questions
Y. Chali, S. Hasan, and K. Imam. Proceedings of the 21st ACM International Conference on Information and Knowledge Management, page 2499--2502. New York, NY, USA, ACM, (2012)
DOI: 10.1145/2396761.2398676
Abstract
This paper addresses the task of answering complex questions using a multi-document summarization approach within a reinforcement learning setting. Given a set of complex questions, a list of relevant documents per question, and the corresponding human-generated summaries (i.e. answers to the questions) as training data, the reinforcement learning module iteratively learns a number of feature weights in order to facilitate the automatic generation of summaries i.e. answers to unseen complex questions. Previous works on this task have utilized a fully automatic reinforcement learning framework that selects the document sentences as the potential candidate (i.e. machine-generated) summary sentences by exploiting a relatedness measure with the available human-written summaries. In this paper, we propose an extension to this model that incorporates user interaction into the reinforcement learner to guide the candidate summary sentence selection process. Experimental results reveal the effectiveness of the user interaction component in the reinforcement learning framework.
%0 Conference Paper
%1 Chali:2012:IPR:2396761.2398676
%A Chali, Yllias
%A Hasan, Sadid A.
%A Imam, Kaisar
%B Proceedings of the 21st ACM International Conference on Information and Knowledge Management
%C New York, NY, USA
%D 2012
%I ACM
%K RNN
%P 2499--2502
%R 10.1145/2396761.2398676
%T Improving the Performance of the Reinforcement Learning Model for Answering Complex Questions
%U http://doi.acm.org/10.1145/2396761.2398676
%X This paper addresses the task of answering complex questions using a multi-document summarization approach within a reinforcement learning setting. Given a set of complex questions, a list of relevant documents per question, and the corresponding human-generated summaries (i.e. answers to the questions) as training data, the reinforcement learning module iteratively learns a number of feature weights in order to facilitate the automatic generation of summaries i.e. answers to unseen complex questions. Previous works on this task have utilized a fully automatic reinforcement learning framework that selects the document sentences as the potential candidate (i.e. machine-generated) summary sentences by exploiting a relatedness measure with the available human-written summaries. In this paper, we propose an extension to this model that incorporates user interaction into the reinforcement learner to guide the candidate summary sentence selection process. Experimental results reveal the effectiveness of the user interaction component in the reinforcement learning framework.
%@ 978-1-4503-1156-4
@inproceedings{Chali:2012:IPR:2396761.2398676,
abstract = {This paper addresses the task of answering complex questions using a multi-document summarization approach within a reinforcement learning setting. Given a set of complex questions, a list of relevant documents per question, and the corresponding human-generated summaries (i.e. answers to the questions) as training data, the reinforcement learning module iteratively learns a number of feature weights in order to facilitate the automatic generation of summaries i.e. answers to unseen complex questions. Previous works on this task have utilized a fully automatic reinforcement learning framework that selects the document sentences as the potential candidate (i.e. machine-generated) summary sentences by exploiting a relatedness measure with the available human-written summaries. In this paper, we propose an extension to this model that incorporates user interaction into the reinforcement learner to guide the candidate summary sentence selection process. Experimental results reveal the effectiveness of the user interaction component in the reinforcement learning framework.},
acmid = {2398676},
added-at = {2018-01-09T17:31:45.000+0100},
address = {New York, NY, USA},
author = {Chali, Yllias and Hasan, Sadid A. and Imam, Kaisar},
biburl = {https://www.bibsonomy.org/bibtex/2e8d1ea017e2a0011675d3cfc1b14f62a/defeatnelly},
booktitle = {Proceedings of the 21st ACM International Conference on Information and Knowledge Management},
doi = {10.1145/2396761.2398676},
interhash = {ea80159933c09f3e8ff9eeb258b71567},
intrahash = {e8d1ea017e2a0011675d3cfc1b14f62a},
isbn = {978-1-4503-1156-4},
keywords = {RNN},
location = {Maui, Hawaii, USA},
numpages = {4},
pages = {2499--2502},
publisher = {ACM},
series = {CIKM '12},
timestamp = {2018-01-09T17:31:45.000+0100},
title = {Improving the Performance of the Reinforcement Learning Model for Answering Complex Questions},
url = {http://doi.acm.org/10.1145/2396761.2398676},
year = 2012
}