Inferring pragmatics from dialogue contexts in simulated virtual agent games

Authors:
Alex Yu-Hung Chien;Von-Wun Soo
Affiliations:
Department of Computer Science, Nationial Tsing Hua University, Hsinchu, Taiwan, R.O.C.;Department of Computer Science, Nationial Tsing Hua University, Hsinchu, Taiwan, R.O.C.,Institute of Information Systems and Applications, Nationial Tsing Hua University, Hsinchu, Taiwan, R.O.C.
Venue:
AEGS'11 Proceedings of the 2011 international conference on Agents for Educational Games and Simulations
Year:
2011

Citing 8
Cited 0

KQML as an agent communication language

CIKM '94 Proceedings of the third international conference on Information and knowledge management
Towards a Better Understanding of Context and Context-Awareness

HUC '99 Proceedings of the 1st international symposium on Handheld and Ubiquitous Computing
Dynamic bayesian networks: representation, inference and learning

Dynamic bayesian networks: representation, inference and learning
Dialogue act modeling for automatic tagging and recognition of conversational speech

Computational Linguistics
Emotive alert: HMM-based emotion detection in voicemail messages

Proceedings of the 10th international conference on Intelligent user interfaces
Identifying agreement and disagreement in conversational speech: use of Bayesian networks to model pragmatic dependencies

ACL '04 Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics
Context-Aware Computing Applications

WMCSA '94 Proceedings of the 1994 First Workshop on Mobile Computing Systems and Applications
Artificial Intelligence: A Modern Approach

Artificial Intelligence: A Modern Approach

Quantified Score

Hi-index	0.00

Visualization

Abstract

Virtual agents in video games may conduct two types of interactions: physical and dialogical. While the former is recognized as gazes and gestures, which received significant attention, the latter is often simplified in simulated virtual agent games. However, dialogical interactions affect the mental states of individual agents, and the relations between them, therefore playing a more important role than physical interactions in games. An implemented dynamic Bayesian Network (DBN) based on speech acts is proposed to model the dialogical effects as dialogue contexts in different aspects, such as emotion states, social relations, and social roles. We adopt a scene in the famous movie Doubt that has 53 dialogue sentences as the test corpus and implement 21 types of speech acts in the experiments. The results indicate that, with our DBN model, agents have the ability of context awareness to infer indirect speech acts from given direct speech acts, and that this ability may assist agents to plan dialogues based on speech acts in future work.