Learning Lexical Alignment Policies for Generating Referring Expressions in Spoken Dialogue Systems Export

@inproceedings{citeulike:4882917, abstract = {We address the problem that different users have different lexical knowledge about problem domains, so that automated dialogue systems need to adapt their generation choices online to the users' domain knowledge as it encounters them. We approach this problem using policy learning in Markov Decision Processes (MDP). In contrast to related work we propose a new statistical user model which incorporates the lexical knowledge of different users. We evaluate this user model by showing that it allows us to learn dialogue policies that automatically adapt their choice of referring expressions online to different users, and that these policies are significantly better than adaptive hand-coded policies for this problem. The learned policies are consistently between 2 and 8 turns shorter than a range of different hand-coded but adaptive baseline lexical alignment policies.}, author = {Janarthanam, Srinivasan and Lemon, Oliver}, booktitle = {Proceedings of the 12th European Workshop on Natural Language Generation}, citeulike-article-id = {4882917}, keywords = {dialogue-management, lexical-alignment, reinforcement-learning}, posted-at = {2009-06-17 19:02:19}, priority = {2}, title = {Learning Lexical Alignment Policies for Generating Referring Expressions in Spoken Dialogue Systems}, year = {2009} }

TY - CONF ID - citeulike:4882917 L3 - citeulike-article-id:4882917 N2 - We address the problem that different users have different lexical knowledge about problem domains, so that automated dialogue systems need to adapt their generation choices online to the users’ domain knowledge as it encounters them. We approach this problem using policy learning in Markov Decision Processes (MDP). In contrast to related work we propose a new statistical user model which incorporates the lexical knowledge of different users. We evaluate this user model by showing that it allows us to learn dialogue policies that automatically adapt their choice of referring expressions online to different users, and that these policies are significantly better than adaptive hand-coded policies for this problem. The learned policies are consistently between 2 and 8 turns shorter than a range of different hand-coded but adaptive baseline lexical alignment policies. TI - Learning Lexical Alignment Policies for Generating Referring Expressions in Spoken Dialogue Systems T2 - Proceedings of the 12th European Workshop on Natural Language Generation KW - dialogue-management KW - lexical-alignment KW - reinforcement-learning AU - Janarthanam, Srinivasan AU - Lemon, Oliver PY - 2009/// ER -