Parameter learning in Bayesian network using semantic constraints of conversational feedback

Authors:
Seung-Hyun Lee;Sungsoo Lim;Sung-Bae Cho
Affiliations:
Dept. Computer Science, Yonsei University, Seoul, Korea;Dept. Computer Science, Yonsei University, Seoul, Korea;Dept. Computer Science, Yonsei University, Seoul, Korea
Venue:
PRICAI'10 Proceedings of the 11th Pacific Rim international conference on Trends in artificial intelligence
Year:
2010

Citing 8
Cited 0

A Bayesian Method for the Induction of Probabilistic Networks from Data

Machine Learning
Learning Bayesian Networks: The Combination of Knowledge and Statistical Data

Machine Learning
Bayesian Networks for Data Mining

Data Mining and Knowledge Discovery
Logic-based subsumption architecture

Artificial Intelligence - Special issue on logical formalizations and commonsense reasoning
Operations for learning with graphical models

Journal of Artificial Intelligence Research
Learning Bayesian network parameters under order constraints

International Journal of Approximate Reasoning
Learning parameters of Bayesian networks from incomplete data via importance sampling

International Journal of Approximate Reasoning
Autonomous Language Development Using Dialogue-Act Templates and Genetic Programming

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

Many learning techniques of Bayesian network have been developed for adaptation to user or environment. However, it seems several drawbacks still exists in conventional learning approach; the hardness of collecting log data, the inherent ambiguity in recognizing and reflecting implicit user's intention, and difficulties in extracting relations between data or definite rules. In this paper, we propose a method for parameter learning in Bayesian network using semantic constraints of conversational feedback to overcome these limitations. Production rules extracted from users' conversational feedback are used in parameter learning of Bayesian network. A comparison test with conventional approaches in conducted to verify the usefulness of the proposed method.