Personalization by Partial Evaluation: A Scenario-Based Approach
Title: Personalization by Partial Evaluation: A Scenario-Based Approach
Abstract: This research explores the use of scenarios to analyze and design personalization systems. It proposes a methodology called PIPE (Personalization is Partial Evaluation), which represents interaction with an information space as a program. This program is then specialized to a user's known interests or information seeking activity by the technique of partial evaluation. The paper demonstrates how designing a PIPE representation can be cast as a search within a space of PIPE models, organized along a partial order. It also shows how personalization can be viewed as the transformation of information spaces to support the explanation of usage scenarios. An example application is described.
Keywords: Personalization, Partial Evaluation, Scenario-Based Design, Explanation-Based Generalization.
Introduction: Personalization is the process of customizing information access to individual users. As the amount of online information grows, the need for personalization becomes increasingly important. This research investigates the use of scenarios to analyze and design personalization systems. It proposes a methodology called PIPE (Personalization is Partial Evaluation), which represents interaction with an information space as a program. This program is then specialized to a user's known interests or information seeking activity by the technique of partial evaluation.
PIPE Modeling: The paper describes how PIPE models can be designed and reasoned about. It presents an example of personalizing a browsing hierarchy and discusses the modeling of these representations. The paper also explores related research in this area.
From Scenarios to Modeling Choices: The paper then focuses on using scenarios to guide the design of PIPE representations. It introduces Explanation-Based Generalization (EBG) and demonstrates how it can be used in personalization. The paper explains how EBG can be used to specialize a theory based on the explanation of an example, and how this can be applied to personalization.
Designing a PIPE Representation: The paper outlines how personalization can be viewed as the transformation of information spaces to support the explanation of usage scenarios. It discusses how this can be achieved by constructing explanations from scenarios and operationalizing these explanations. The paper also presents an example application of PIPE.
Discussion: The paper concludes with a discussion of the benefits and challenges of using scenarios to design personalization systems. It also highlights the potential of PIPE as a framework for personalization research.
Appendix: An Example Application: The paper concludes with an example application of PIPE, which demonstrates how the methodology can be applied to personalize information access.
Conclusion: In conclusion, this research explores the use of scenarios to analyze and design personalization systems. It proposes a methodology called PIPE (Personalization is Partial Evaluation), which represents interaction with an information space as a program. This program is then specialized to a user's known interests or information seeking activity by the technique of partial evaluation. The paper demonstrates how designing a PIPE representation can be cast as a search within a space of PIPE models, organized along a partial order. It also shows how personalization can be viewed as the transformation of information spaces to support the explanation of usage scenarios. An example application is described.
Link to Article: https://arxiv.org/abs/0111007v1 Authors: arXiv ID: 0111007v1