Read more
As the range of people gaining access to information is widening a custom-made presentation of information is growing more and more important. This book presents an approach that allows a multimedia presentation system to cater for the users' needs. Susanne van Mulken discusses the relevant literature on cognitive psychology, current multimedia presentation systems, and user modeling, and she develops a system module which - on the basis of Bayesian networks - is able to estimate the decodability of planned presentations for an individual user. The module learns facts about the user by means of the interaction and diagnoses the critical parts of a display so that they can be remedied before the actual presentation. Selected models are empirically validated by psychological experiments.
List of contents
1 Introduction.- 1.1 Multimedia Dialog Systems.- 1.2 The Presentation System PPP.- 1.3 Some Terminology.- 1.4 Focus and Goals of this Thesis.- 1.5 Overview of the Thesis.- 2 Comprehension in Multimedia Communication.- 2.1 Optimization of Mental Models.- 2.2 Relevant Factors for Comprehension.- 2.3 Two Modes of Comprehension.- 2.4 Comprehension of Text and Graphics.- 2.5 Comprehension of Technical Instructions.- 2.6 Conclusion.- 3 User Characteristics in Current Presentation Systems.- 3.1 Intelligent Multimedia Presentation Systems.- 3.2 Content Selection and Organization.- 3.3 Media Selection.- 3.4 Subcode selection.- 3.5 Generation of Referring Expressions.- 3.6 Layout.- 3.7 Conclusion.- 4 User Modeling: Representation and Inference.- 4.1 Purpose and Terminology.- 4.2 Dimensions of User Modeling.- 4.3 The Contents of the User Model.- 4.4 Acquisition of User Models.- 4.5 Exploitation of User Models.- 4.6 Representation and Inference.- 4.7 Conclusion.- 5 Modeling Decoding Problems.- 5.1 Representation of Inferences in Ppp.- 5.2 Partial Evaluations of Displays.- 5.3 Overall Evaluation of Displays.- 5.4 Feedback to Ppp.- 5.5 Implementation.- 5.6 Conclusion.- 6 Empirical Studies.- 6.1 Experiment I: Object Identification.- 6.2 Experiment II: Meaning Decoding.- 6.3 Conclusion.- 7 Achievements and Future Work.- 7.1 Scientific Contributions.- 7.2 Limitations.- 7.3 Possible Extensions.- 7.4 Application in Future Presentation Systems.- A Bayesian Networks: Belief Update and Belief Propagation.- A.1 Belief Update.- A.2 Belief Propagation.- A.2.1 Upward Propagation.- A.2.2 Downward Propagation.- B Example Network.- C Instructions Used in the Experiments.- C.1 Experiment I.- C.2 Experiment II.- D Materials Used in the Experiments.- D.1 Experiment I.- D.2 Experiment II.- E Scatterplots of Correlations.- References.
About the author
Dr. Susanne van Mulken ist Forschungsmitarbeiterin am Deutschen Forschungszentrum für Künstliche Intelligenz GmbH in Saarbrücken. Sie promovierte 1998 bei Professor Dr. Werner H. Tack an der Universität des Saarlandes.
