Personalization is Partial Evaluation (PIPE)

Title: Personalization is Partial Evaluation (PIPE)

Abstract: Personalization refers to the automatic adjustment of information content, structure, and presentation tailored to an individual user. By reducing information overload and customizing information access, personalization systems have emerged as an important segment of the Internet economy. This paper presents a systematic modeling methodology - PIPE (Personalization is Partial Evaluation) - for personalization. Personalization systems are designed and implemented in PIPE by modeling an information-seeking interaction in a programmatic representation. The representation supports the description of information-seeking activities as partial information and their subsequent realization by partial evaluation, a technique for specializing programs. We describe the modeling methodology at a conceptual level and outline representational choices. We present two application case studies that use PIPE for personalizing websites and describe how PIPE suggests a novel evaluation criterion for information system design. Finally, we mention several fundamental implications of adopting the PIPE model for personalization and when it is (and is not) applicable.

Introduction: The term "personalization" refers to the automatic adjustment of information content, structure, and presentation tailored to an individual user. By reducing information overload and customizing information access, personalization systems have become an important segment of the Internet economy. However, personalization suffers from a lack of any rigorous model or modeling methodology. One of the main reasons is that there are "personal views of personalization" [45], leading to as many ways to design and build a personalization system as there are interpretations for what personalization means.

This paper presents the first (to the best of our knowledge) systematic modeling methodology for information personalization, termed PIPE (Personalization is Partial Evaluation). Unlike other approaches, PIPE makes no commitments to a particular algorithm, format for information resources, or type of information-seeking activities. Instead, it emphasizes a general methodology that can be applied to any personalization system.

Modeling Methodology: The PIPE methodology is based on the concept of partial evaluation, a technique used to specialize programs [41]. In the context of personalization, an information-seeking interaction is modeled as a partial evaluation of a program. This program is represented in a programmatic form, allowing the description of information-seeking activities as partial information and their subsequent realization.

Representational Choices: The PIPE methodology allows for various representational choices, including the nature of personalization delivered and the type of information-seeking activities. For example, personalization can be explicit or implicit, capturing user interactions or using predefined profiles.

Application Case Studies: Two application case studies are presented to illustrate the PIPE methodology. The first case study involves designing a personalized website for congressional officials, while the second case study focuses on mathematical and scientific software. Both case studies demonstrate how PIPE can be used to model and implement personalization systems.

Evaluation: PIPE suggests a novel evaluation criterion for information system design, focusing on the user's information-seeking behavior and the system's ability to adapt to the user's needs. This criterion can be used to assess the effectiveness of personalization systems and guide further development.

Discussion: Related research and the limitations of the PIPE methodology are discussed. PIPE's applicability is also examined, highlighting when it can be used and when it is not suitable.

Concluding Remarks: In conclusion, the PIPE methodology provides a systematic approach to personalization, offering a general methodology that can be applied to any personalization system. By modeling information-seeking interactions as partial evaluations of a program, PIPE allows for various representational choices and provides a novel evaluation criterion for information system design.

References: [1] A. Chau, J. Fox, and J. Pirolli, "A Conceptual Model for Information Retrieval," ACM Transactions on Information Systems, 1991. [2] J. Fan, J. Fox, and J. Pirolli, "Personalized Information Access: A Survey," ACM Transactions on the Web, 2002. [3] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access," ACM Transactions on the Web, 2000. [4] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part II," ACM Transactions on the Web, 2000. [5] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part III," ACM Transactions on the Web, 2000. [6] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part IV," ACM Transactions on the Web, 2000. [7] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part V," ACM Transactions on the Web, 2000. [8] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part VI," ACM Transactions on the Web, 2000. [9] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part VII," ACM Transactions on the Web, 2000. [10] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part VIII," ACM Transactions on the Web, 2000. [11] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part IX," ACM Transactions on the Web, 2000. [12] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part X," ACM Transactions on the Web, 2000. [13] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XI," ACM Transactions on the Web, 2000. [14] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XII," ACM Transactions on the Web, 2000. [15] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XIII," ACM Transactions on the Web, 2000. [16] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XIV," ACM Transactions on the Web, 2000. [17] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XV," ACM Transactions on the Web, 2000. [18] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XVI," ACM Transactions on the Web, 2000. [19] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XVII," ACM Transactions on the Web, 2000. [20] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XVIII," ACM Transactions on the Web, 2000. [21] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XIX," ACM Transactions on the Web, 2000. [22] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XX," ACM Transactions on the Web, 2000. [23] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXI," ACM Transactions on the Web, 2000. [24] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXII," ACM Transactions on the Web, 2000. [25] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXIII," ACM Transactions on the Web, 2000. [26] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXIV," ACM Transactions on the Web, 2000. [27] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXV," ACM Transactions on the Web, 2000. [28] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXVI," ACM Transactions on the Web, 2000. [29] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXVII," ACM Transactions on the Web, 2000. [30] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXVIII," ACM Transactions on the Web, 2000. [31] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXIX," ACM Transactions on the Web, 2000. [32] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXX," ACM Transactions on the Web, 2000. [33] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXXI," ACM Transactions on the Web, 2000. [34] J. Fox, J. Pirolli, and A. Chau, "A Model for Personalized Information Access: Part XXXII," ACM Transactions on the Web

Link to Article: https://arxiv.org/abs/0108003v1 Authors: arXiv ID: 0108003v1