System Architecture for Psychological Customization of ... - CiteSeerX

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

System Architecture for Psychological Customization of Communication Technology Marko Turpeinen Helsinki Institute for Information Technology [email protected] Timo Saari M.I.N.D. Lab Center for Knowledge and Innovation Research, Helsinki School of Economics [email protected]

Abstract Personalization is a process that changes the functionality, interface, information content, or appearance of a system to increase its personal relevance to an individual. Personalization systems accommodate individual’s needs and interests explicitly through changes and selections initiated by the user, and implicitly through automatic adaptation techniques. Currently most of the emphasis in personalization systems is geared towards the utilitarian aspects of personalized information delivery. However, what is lacking is the customization of information based on its likely emotional and cognitive effects on different users of communication technology. Information presented to individual users or a group of users may be customized on the basis of the immediate emotional and cognitive types of psychological effects it is likely to enable or create in certain individuals or groups. Both content and its way of presentation (modality, visual layouts, ways of interaction, structure) may be varied. Despite obvious complexities empirical evidence suggests that the way of presenting information to certain psychological profiles has predictable psychological effects. For instance, one may facilitate positive emotion for users with certain personality type or more efficient learning for certain cognitive styles. This is the basic concept of Mind-Based Technologies. Psychological Customization may be considered an operationalization and technique of implementing the concept of Mind-Based Technologies in system design. Psychological customization may be used for controlling social interaction-centric and individual-

centric influence of the content and way of presentation of information on consequent and transient emotional and cognitive effects. This paper explores the design space and presents a basic system architecture to implement Psychological Customization.

1. Introduction Digital media allow for unprecedented flexibility in content creation, access, interactivity and customization. A wide variety of Internet-based technologies for content management and personalization have been developed to better serve the needs of individuals and communities. For example, news media can now be customized by filtering a subset of available news stories for the reader, or by augmenting selected news stories with personally relevant related material or illustrative comparisons as stipulated by the user. Digital communication systems may be considered as consisting of three layers [1]. At the bottom resides a physical layer that includes the physical technological device and the connection channel that is used to transmit communication signals. In the middle is a code layer that consists of the protocols and software that make the physical layer run. At the top is a content layer, which consists of multimodal information. The content layer includes both the substance and the form of multimedia content [2, 28]. Substance refers to the core message of the information. Form implies aesthetic and expressive ways of organizing the substance, such as using different modalities and structures of information [28].

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

1


These communication systems are amenable for various techniques of customization. The content can be selected, grouped, and organized, form can be tailored to suit the needs and preferences of the individual, and delivery methods can be tailored by media platform capabilities, update time and frequency, and cost. Customization is individualcentric when the content or service is personalized according to a user profile, or more community-centric when the content or service is customized for a group of users based their shared on a community profile [31]. Personalization is a process that changes the functionality, interface, information content, or appearance of a system to increase its personal relevance to an individual. Personalization systems accommodate individual’s needs and interests explicitly through changes and selections initiated by the user, and implicitly through automatic adaptation techniques. These two approaches are often used to complement each other in personalized services. Explicit profiling is the simplest method to produce a user interest profile. However, this method requires the user to go through and learn additional steps of functionality typically at the early stages of usage when the user is not at all familiar with the system. With implicit feedback the system collects user interests indirectly by monitoring the interaction between the user and the application. Learning algorithms make the profile adapt over time more closely to the viewer’s habits. Currently most of the emphasis in personalization systems is geared towards the utilitarian aspects of personalized information delivery. However, it seems that other dimensions of customization technology are increasingly popular and equally important from the point-of-view of user experience. Users are also often active in tailoring the technological devices and systems to better reflect their own personality and identity. Selecting a new ring tone for a mobile phone, or changing the outlook (i.e. the “skin”) of one’s favorite media player, are popular examples of customization that typically reflect the psychological profile of the user. These forms of user-initiated personalization of the appearance of electronic products have largely been overlooked in literature [3]. Personalization systems may also be conceptualized as Mind-Based. Technologies may be considered Mind-Based because they systematically take into account the characteristics and individual differences of different segments of users and alter information presented to them in a systematic manner to create emotional and cognitive effects [28]. Psychological Customization may be considered an operationalization

and technique of implementing the concept of MindBased Technologies in system design. This paper explores the role of psychological customization technologies in designing and delivering personalized content, applications and services to the users. The concept of psychological customization is used to especially highlight the importance of individual differences regarding immediate emotional and cognitive effects of information technology in their individual users. Then a system architecture is presented that provides a framework for developing these “Mind-based” customization technologies.

1.1 Related work Key technical challenges in the implementation of personalization systems typically include (1) semantically rich modeling of content and services using structured metadata, (2) modeling the changing short-term and long-term profiles of the users, and (3) providing suitable system architecture for the personalization system [9]. It is relatively simple to use existing content management systems to personalize the appearance of content, such as layout, colors, and fonts of a Web site, in response to the user’s preferences. For example, My Excite (http://www.excite.com) has a selection screen for themes and color schemes to be used on this highly personal site (Figure 1).

Figure 1. Theme selection on My Excite. Psychological customization differs from this userinitiated selection of Web site’s outlook. This paper proposes the approach of modeling the user’s psychological profile (such as aesthetic preferences, cognitive style and personality) at a higher level of abstraction and then using this profile data to

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

2


systematically provide user experiences that best match these higher level user characteristics. For instance one may vary the form of information per user profile, which systematically produces, amplifies, or shades different psychological effects [28]. The need for structured metadata is motivated by research that states that domain models are central to understanding, and especially hierarchical categorizations are important in making the connections between the meaning of the media content and the concepts understood by an individual [30]. Therefore, there is a need for an explicit domain model, which is used when deep multi-dimensional metadata is created for customized media services. MIT Media Lab’s fishWrap [4] was one of the first prototypes for personalized newspapers using structured metadata and user profiling. Currently there is a multitude of similar services that provide news filtering for individuals at different levels of sophistication (e.g. see [18], for a detailed illustration on My Yahoo’s personalization features). Future steps in ontology development and ontology-based user profiling are strongly linked to the current approaches taken in the work related to the Semantic Web. The development of rich machine-useable semantic description of content is not at the core of psychological customization, as in most cases, it is not tied to a specific content domain. On the contrary, the psychological customization system should maximize the applicability of user’s psychological profile data across multiple content domains and applications. Kobsa [13] introduces three dimensions of profile data that can be acquired. He refers to these as user data (personal characteristics of the user), usage data (user’s interactive behavior), and environmental data (the nature of the user’s system). Psychological customization systems are mostly based on user data, but can also benefit from usage and environmental profiles. There is on-going work on specifications related to these different types of profile data, such as CC/PP and UAProf for environmental data, that both focus on device-independent Web access and the technical capabilities of terminal devices. Recommender systems using collaborative information filtering techniques categorize users automatically into neighborhoods based on similarities between user profiles. The tools use these neighborhoods to recommend new items to similar users, or to recommend users to each other [24]. LifestyleFinder was an early example of a recommender system that was not based on user’s interests, but rather their overall lifestyles [12]. Similarly, psychological customization systems can base the recommendations on user’s psychological profile data.

Fogg [7] emphasizes the potential of “captology” that include system design and electronic devices that systematically apply the principles of persuasion. He also describes the “Kairos principle” of interfering or offering suggestions at opportune moments. Fogg also discusses in length the ethical issues regarding persuasion technology. These themes are directly related to the underlying design rules and principles applied in psychological customization systems as described in this paper. Psychological customization as a research area has some overlap with usability studies and design studies. For instance, in usability studies the pleasantness and the aspect of having fun with interfaces have been addressed [20]. Affective computing has been developed in the area of computers and emotion [21]. Accordingly, in design-related research there has been discussion about emotion and design [8]. However, according to the authors’ knowledge no other comprehensive framework of varying form of information to systematically create emotional and cognitive effects has been presented.

2. Psychological customization When perceiving information via media and communications technologies, the mind is psychologically transported into a quasi-natural experience of the events described. This is called presence. In presence, information becomes the focused object of perception, while the immediate, external context, including the technological device, fades into the background [17]. Various empirical studies show that information experienced in presence has real psychological effects on perceivers, such emotion based on the events described or cognition of making sense of the events and learning about them [23]. When using collaborative technology for computer mediated social interaction, users experience a state called social presence during which users may for instance experience intimacy of interaction or feeling of togetherness in virtual space [17]. During social presence users also experience various other types of emotional and cognitive effects, such as interpersonal emotion, emotion based on being successful at the task at hand and learning from shared activities or shared information. The social dimension of presence and its cognitive and emotional implications have not been sufficiently researched, however. When a user is interpreting information, a complex set of interrelated “gateway variables” may influence the “outcome variables” like his psychological states, such as presence, and experience of the information,

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

3


including learning and emotion. These gateway variables may be clustered as mind (individual differences and social similarities of perceivers), content (information substance and form embedded in technology with certain ways of interaction) and context (social and physical context of reception) [27, 28]. Being able to systematically and reliably predict the relationships of the gateway variables is the key to mind-based media and communications technologies and psychological customization. This is complex and it may be sensible to concentrate only on the most robust psychological effects with certain tasks. With the possibility of real-time customization and adaptation of information for different perceivers it is hypothesized that one may vary the form of information within some limits per the same substance of information. For instance, the same substance can be expressed in different modalities, or with different ways of interaction with the user and technology. This may produce a certain psychological effect in some perceivers; or shade or amplify a certain effect. This approach may also be suitable for creating psychological effects during computer mediated social interaction. Traditional mass media channels have not been amenable to efficient customization. At best, these media have been able to tailor content to reflect perspectives of a local community within the parameters used in market segmentation strategies. Internet-based technologies for content management and personalization introduce a new set of tools for serving individuals and communities of much wider variation sizes and types [6, 26]. Based on the principle of variability, many potential versions of the same media product or a particular collection of information may be available for different perceivers [19]. One may discuss the “packaging” of information, which means how the different dimensions of information are put together into a certain type of package, including form and substance. The content can be selected and organized in different ways and the presentation of content can be tailored to suit the needs and preferences of the individual. This may include personal preferences for layouts or color schemes. This tailoring may result in different looking products based on the display device or publication style [32].

Layer of technology Physical

Code

Content

Key factors Hardware - large or small vs. human scale - mobile or immobile - close or far from body (intimatepersonal-social distance) Interaction - degree of user vs. system control and proactivity through user interface Visual-functional aspects - way of presenting controls in an interface visually and functionally Substance - the essence of the event described - type of substance (factual/imaginary; genre, other) - narrative techniques used by authors Form 1. Modalities - text, video, audio, graphics, animation, etc.

2. Visual layout - ways of presenting various shapes, colours, font types, groupings and other relationships or expressive properties of visual representations - ways of integrating modalities into the user interface

3. Structure - ways of presenting modalities, visual layout and other elements of form and their relationships over time - linear and/or non-linear structure (sequential vs. parallel; narrative techniques, hypertextuality)

Table 1. Key factors influencing individualcentric emotional and cognitive effects of technology, adapted from [28].

Table 1 addresses the key factors, which may influence psychological effects of information. For instance, a user may wish to have certain substance of information with as much video material as possible and have the system completely take over the control and present a personalized tv-newscast-type of information flow. Another example would be that the user has a profile that indicates it is beneficial for him to receive information in textual modality and the system may try to alter the information flows presented to him accordingly. The view of the authors of the paper is that psychological effects occurring during social presence

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

4


when in social computer mediated interaction have not been sufficiently researched. It can be hypothesized, however, that at least roughly similar psychological influences can be created via personalizing substance and form of information during social interaction as with individual-centric human-computer interaction. Thus the idea of gateway variables could be extended to tasks involving social interaction in the light of the factors presented in Table 1. The relevance of this framework to personalization research and HCI is evident as it may provide an approach to gaining access to and partly controlling the psychological states of the users of information systems.

2.1 Validating the need for psychological customization Even though no actual system has been implemented yet for psychological customization, empirical evidence supports the feasibility and validity of the idea of psychological customization. The key idea is that there seem to be several cognitive and emotional effects that are moderated by individual differences, for instance. This suggests the need for psychological customization systems that optimize the presentation of information to different target groups having different psychological profiles. For instance, Ravaja [22] examined the moderating influence of dispositional Behavioral Inhibition System (BIS) and Behavioral Activation System (BAS) sensitivities, Negative Affect, and Positive Affect on the relationship between a small moving vs. static facial image and autonomic responses when viewing/listening to news messages read by a newscaster with 36 subjects. Autonomic parameters measured were respiratory sinus arrhythmia (RSA), low-frequency component of heart rate variability, electrodermal activity (EDA), and pulse transit time (PTT). The results showed that dispositional BAS sensitivity, particularly BAS Fun Seeking, and Negative Affect interacted with facial image motion in predicting autonomic nervous system activity. A moving facial image was related to lower RSA and EDA and shorter PTTs as compared to a static facial image among high fun seekers, while there was no, or an inverse, relationship between these variables among low fun seekers. Facial image motion may contribute to sustained attention particularly among high fun seekers, given that it may increase the so-called sense of presence and act as a positive incentive for high fun seekers, partly because of their higher need for stimulation.

Also, there is ample evidence in literature that varying the form of information creates emotional and cognitive effects. In media studies it has been found that different modalities, such as visual and auditory, may lead to different kinds of psychological influences and the valence of a preceding subliminal stimulus influences the subsequent evaluation of a person evaluated [5, 11]. In educational studies it has been shown that different ways of processing information influence learning and emotion of stimuli with certain modality [25]. Research concerning emotional influences on the cognitive processing of information has often concentrated on how different emotions related to information change the way users pay attention to, evaluate and remember the mediated message. This research has results on the influence of emotional information as increasing the user’s selfreported emotion [16]; attention (physiological and self-reported) [15] and memory for mediated messages, particularly arousing messages [14, 15]. Studies in experimental psychology have shown that recognition and memory can be influenced or even enhanced by previous exposure to subliminal visual or auditory images of which the subjects are not consciously aware [10]. Some of these effects are produced in interaction with individual differences, such as cognitive style, personality, age and gender.

2.2 Psychological customization for groups and communities The idea of Mind-Based Technologies in producing psychological effects for both individuals and groups may be utilized with Psychological Customization. For instance, if one wishes to produce more or less emotion with certain form of information embedded in a particular device with a certain user interface, one would have to know which types of variations of form may cause which types of qualities of emotion for the different perceivers. The same principle may apply to other psychological effects, such as presence, learning, persuasion. Emotional and cognitive effects of information are related to communication within social networks as follows: one may manipulate manually or with automated systems the substance and the form of information. It is obvious that in social interaction the users may construct the substance and form of the messages exchanged manually. However, for instance the form of the message may be varied with automated systems. The information needed to conduct these automated manipulations can be accessed via individual and social modeling and profiling of the users. For instance, a user with an intention to create

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

5


positive emotion in the other user with the way of presenting his textual message may utilize a background color predicted to induce a mild positive emotional state in the receiving user based on his profile.

3. Design space customization

for

psychological

Figure 2 presents a taxonomy of the design space for application areas of psychological customization. Application areas are arranged according to (1) whether they are more focused on individual-centric human-computer interaction or computer mediated social interaction, and (2) whether they are more related to media content or messaging-oriented (including both human-human and human-computer dialogue). It is hypothesized that the selection and manipulation of substance of information takes place through the technologies of the application areas. Underlying the application areas is a basic technology layer for customizing design. This implies that within some limits one may automatically vary the form of information per a certain category of substance of information. The design space for psychological customization is formed in the interaction of a particular application area and the possibilities of the technical implementation of automated design variation. For instance, Augmentation System refers to a system, which may enhance cognitive processing and understanding of a particular substance of information. In news services, a particular piece of information may be surrounded by articles related to the base-article selected on the basis of a user profile. Displaying related articles or other additional information to the user may enhance the understanding of the basic story [31]. Augmentation may be varied for instance for experts and novices. Adding the possibilities of realtime variation of design may produce applications in the both the substance and form of information is varied for maximum cognitive efficiency. This means that the augmentations may be altered by modality or some other means per user profile. Information Filtering refers to a system in which a software program filters the substance of information according to the user profile. Also here one may automatically vary the design within some limits per user. Notification System implies a way for the computer to alert the user of some noteworthy events, such as arrival of new email. It may be that also here one may vary the design of the substance of such a notification

to create an effect of urgency or pleasantness, for instance. CONTENTINTE NS IVE

Augmentation System

Collaborative Filtering

Information Filtering

Multi-player Online Game

Affective Computing INDIVIDUAL

SOCIAL

Persuasive Technology

Notification System

Multimedia Messaging System

Text Messaging System

MESSAGINGINTE NS IVE

Figure 2. Design space for Psychological Customization

Affective Computing refers to systems in which the computer is receiving real-time feedback of the emotional state of the user and may use this information to adapt its actions, such as when displaying certain substance of entertainment information, like a movie. With automated design variation, one may vary some emotional components of the substance to create desired emotional effects. A feedback system may then pick up the efficiency of the variations made to make sure an effect has been realized in a particular user. Collaborative Filtering is a technique to offer personalized substance of information or recommendations based on clustering individual users into groups by buying behavior or some other dimension. Collaborative Filtering as a technique may be altered by automated variation of design to present the substance, for instance a book recommendation as substance may be altered in design to create a desired effect, such as maximum credibility of the recommendation for a particular user. Persuasive Technology refers to human-computer interaction in which there is an underlying goal to change the attitudes and behavior of the user [7]. For instance, one may motivate users to quit smoking via motivating games. Also socially intelligent agents may

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

6


be classified as persuasive technology. Often with agents an illusion of being in interaction with another human being is created in the user via using for example animated agents in e-Commerce. It is known that both the substance of the interaction (what is being sold and what the agent says) and the form of interaction (how information is presented, what is the appearance and personality of the agent) influence for instance trust, persuasiveness, emotion and liking of the transaction [23]. What psychological customization may add here may be more personalization of the way of presenting information in an eCommerce site as well as it may vary the appearance or other features of the agent without changing the substance, i.e. what the agent says. A Messaging System entails computer mediated social interaction, like SMS, MMS, chat or CSCWsoftware usage. By varying the design of MMS messages for instance one may be able to create emotional effects in the receivers of the message. Similarly, cognitive processes may be optimized for each individual user with the personalization of different aspects of the form of information embedded in a CSCW-software suite.

4. System architecture for psychological customization Psychological customization requires a software infrastructure that models a user, or a user group, across the boundaries of individual applications or services. In addition, a database of design rules is needed to define the desired cognitive and emotional effects for different types of profiles. Once these components are in place, existing content management technologies can be extended to cover variations of form and substance of information based on psychological profiles and design rules to create the desired psychological effects. This system architecture is presented in Figure 3. Psychological Customization System (PCS) is a new form of middleware between applications, services, content management systems and databases. It provides an interface for designing desired psychological effects and user experiences for individual users or user groups. In essence, Psychological Customisation System goes beyond the technical or task-based approach to content management technologies. It poses the question: how to utilize these technologies to create desired optimal psychological effects, such as positive emotion and efficiency of cognition with certain substance or within certain collaborative tasks across applications and services?

USER

SERVICE

APPLICATION APPLICATION APPLICATION

PSYCHOLOGICAL CUSTOMIZATION SYSTEM

SERVICE SERVICE

CONTENT MANAGEMENT CONTENT DATABASE

RULES AND SCENARIOS MIDDLEWARE OPERATING SYSTEM HARDWARE

Figure 3. High-level architecture psychological customization system

for

Personalization can be system and/or user-initiated. User modeling covers static and dynamic aspects to model personal preferences, rights, experiences, and behaviors. Psychological customization system provides support for dynamic profile retrieval, adaptation and management personalization includes service adaptation to the specific usage scenarios, content adaptation to terminal capabilities, ambient information and content filtering on the basis of user's preferences. To be able to conduct psychological customization the system needs a model of the user or users. In fact, customization is always based on some type of model of an individual, a group or a community. These three can be considered separately: • user modeling, which includes a profile of an individual user, • user clustering, which is based on similarities between user profiles and forming a user cluster using some form of automated technique, and • community modeling, which includes a profile about the social group as a whole, not as the sum or the average of its individual member’s profiles. A user model is computer-accessible presentation of information about an individual regarding specified domains of use. This user model can consist of data explicitly given by the user for the purposes of modeling, implicit observations of user’s behavior. The personalization can also be based on inferred data about the user that is stored in the user model. The profile data representing the user model physically

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

7


reside on the user’s device or on the server that hosts the psychological customization system, or on both. Community modeling can be used to model the collective group, especially in a case of joint activity. User modeling, user clustering and community modeling may be used simultaneously in customization [31]. Our approach is to promote this hybrid usage of individual and social modeling in the context of psychological customization. Privacy is one of the main social and technical challenges underlying any system that customizes media content for individual users. Considering the intimacy of psychological effects and user profiles the role of privacy protection is essential. The privacy issues in personalization are dependent on the physical storage location and modes of access to personal information. In the area of Web-based applications, the PCS can be seen as an extension to existing Web services and application server frameworks, such as Microsoft’s .Net or Java 2 Enterprise Edition (J2EE). However, the presented high-level system architecture can also be applied to other domains, such as games.

4.1 Example implementation using J2EE The most popular framework for building customized Web-based applications is Java 2 Enterprise Edition. J2EE-based implementation of the Psychological Customization System for Web-based applications is depicted in Figure 4. The basic J2EE three-tiered architecture consisting of databases, application servers, and presentation servers has been extended with three middleware layers: content management layer, customer relationship management layer, and psychological customization layer. The profiles of the users and the communities are available in the profile repository. The Content Management System is used to define and manage the content repositories. This typically is based on metadata descriptions of the content assets. The metadata of the content repositories is matched against the user and community profiles by the Customer Relationship Management (CRM) system. The CRM system includes tools for managing the logic behind content, application and service customization. Rules can be simple matching rules or more complex rule sets. A special case of a rule set are scenarios, which are rule sets involving sequences of the interactions on the Web site. The Customer Relationship Management layer also includes functionality for user and community modeling. This layer can also perform automated customer data analysis, such as user clustering.

User User

User profile

User

Community

Community profile

Web Presentation Layer Profiles Psychological Customization System Rules and Scenarios

Customer Relationship Management

Content

Content Management System Application Server

Figure 4. J2EE implementation of Psychological Customization System

the

The Psychological Customization System layer performs the optimization of the form of the content as selected by the Customer Relationship Management layer. This functionality can be considered similar to the device adaptation by using content transformation rules (for example XSL-T). In the case of the psychological customization, the transformation rules are produced based on the design rules for content presentation variation and the contents of the psychological profile of the user. After this optimization, the content is passed to the Web presentation layer.

5. Conclusion This paper has introduced the idea of Mind-Based Technologies and Psychological Customization. The key ideas of the paper have been contextualized in mainstream personalization and customization research. However, according to the authors’ knowledge no other comprehensive framework of varying form of information to systematically create emotional and cognitive effects has been presented. Differences to other approaches to influencing user experience of information systems are various. Usability studies traditionally address the question of

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

8


how to make difficult technology easy to use. Usability is at least partly founded on the idea of optimal humanmachine performance, i.e. how well a user can manipulate and control a machine. While this is certainly important, in consumer applications, like games, online news, e-Commerce and peer-to-peer computer mediated social interaction the emphasis of the use of the computer may well be in the arena of psychological effects rather than on usability as such. Games may be used to experience emotional arousal and excitement, online news may be used to learn of the events of the world and eCommerce vendors present their wares in a persuasive manner to users. Consequently, Psychological Customization is founded on the idea of creating a desired psychological effect with the available means of automatic variation of substance and form of information. It may be seen as partly based on or adding to usability studies. Naturally, if an interface is not usable, it may not be possible to systematically create psychological effects. Design-based approaches to interface design have adopted the perspective of creating desired experiences, such as positive emotion for a user. However, what is lacking is the systematic and explicit, communicable, knowledge of what exactly in the elements of design may produce such effects. Also, the influence of individual differences of the users in the variation of the effects remains unknown. Moreover, it may be difficult to alter hand-made designs rapidly, and almost impossible to do it in realtime. Psychological Customization does not claim to replace design as such, it may rather be a tool for designers to systematically vary some elements of an interface in real-time within a hand-designed template, for example. From the point of view of theoretical contribution, Psychological Customization poses a possible change in the perspective to technology. One may view technology as a source for creating added value for a user, such as enabling desired emotional and cognitive effects. To be able to realize Psychological Customization one may have to conduct a number of experimental studies in which certain applications with certain tasks are tested in laboratory and field conditions. Laboratory methods, such as psychophysiology, are be used for indicating emotional effects in addition to qualitative methods. The architecture presented in this paper leaves open many related implementation issues of the Psychological Customization System. The future goals of this research include building prototypes that use the design principles outlined in this paper, especially in the area of mobile devices, persuasive technology and games. The prototype implementation may then be

used for further exploring and studying the design space of Psychological Customization as well as for conducting field studies on the emerging cognitive and emotional effects in different types of users of the system.

6. References [1] Y. Benkler, “From Consumers to Users: Shifting the Deeper Structures of Regulation.” Federal Communications Law Journal, 52, 2000, pp. 561-63. [2] D. Billmann, “Representations.”, in Bechtel, W. and Graham, G. A companion to cognitive science, Blackwell publishers, Malden, MA, 1998, pp. 649-659. [3] J. Blom, and A. Monk, “Theory of personalisation of appearance: why people personalise their mobile phones and PCs.”, to appear in Journal of Human-Computer Interaction. [4] P. Chesnais, M. Mucklo, and J. Sheena, “The Fishwrap Personalized News System”. Proceedings of the 2nd International Workshop on Community Networking, Princeton, NJ, 1995. [5] R. Cuperfain and T. K. Clarke, “A new perspective on subliminal perception”, Journal of Advertising, 14, 1985, pp. 36-41. [6] T. Erickson, “Designing agents as if people mattered”, in Bradshaw, J. M. (eds.) Software agents, AAAI Press/The MIT Press, Menlo Park, CA, 1995, pp. 79-96. [7] B. J. Fogg, Persuasive technology. Using computers to change what we think and do, Morgan Kaufmann Publishers, New York, 2002. [8] H. Hirsh, C. Basu and B. Davison, “Learning to personalize”, Communications of the ACM, 43(8), 2000, pp.102-106. [9] S. Jokela, M. Turpeinen, T. Kurki, E. Savia, and R. Sulonen, “The Role of Structured Content in a Personalized News Service”, Proceedings of the HICSS-34, IEEE Computer Society, Maui, Hawaii, USA, January 3-6, 2001. [10] J. F. Kihlström, T. M. Barnhardt and D. J. Tataryn, ”Implicit perception”, in R. F. Bornstein and T. S. Pittmann, (Eds.), Perception without awareness. Cognitive, clinical and social perspectives, Guilford Press, New York, 1992, pp. 1754. [11] J. A. Krosnick, A. L. Betz, J. L. Jussim, and A. R. Lynn,. "Subliminal conditioning of attitudes", Personality and Social Psychology Bulletin, 18, 1992, pp. 152-162. [12] B. Krulwich, "LifestyleFinder: Intelligent user profiling using large-scale demographic data", AI Magazine, vol. 18, no. 2, 1997, pp. 37-45.

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

9


[13] A. Kobsa, A. Koenemann, and W. Pohl, Personalized hypermedia presentation techniques for improving online customer relationships, Sankt Augustin: GMD Forschungszentrun Informationstechnik GmbH, 1999 [14] A. Lang, "Involuntary attention and physiological arousal evoked by structural features and mild emotion in TV commercials", Communication Research, 17 (3), 1990, pp. 275-299. [15] A. Lang, P. Dhillon, and Q. Dong, "Arousal, emotion and memory for television messages", Journal of Broadcasting and Electronic Media, 38, 1995, pp. 1-15. [16] A. Lang, J. Newhagen, and B. Reeves, "Negative video as structure: Emotion, attention, capacity and memory", Journal of Broadcasting and Electronic Media, 40, 1996, pp. 460-477. [17] M. Lombard and T. Ditton, "At the heart of it all: The concept of presence", Journal of Computer Mediated Communication, 3 (2), 1997. [18] U. Manber, A. Patel and J. Robison, "Experience with personalization on Yahoo", Communications of the ACM, 43(8), 2000, pp. 35-39. [19] L. Manovich, The language of new media, MIT Press, Cambridge, MA, London, England, 2001. [20] A.F. Monk and D. Frohlich, "Computers and Fun", Personal Technology, 3, 1999, p. 91. [21] R. W. Picard, Affective Computing. MIT Press, Cambridge, 1997. [22] N. Ravaja, "Presence-related influences of a small talking facial image on psychophysiological measures of emotion and attention", Proceedings of the 5th Annual International Workshop Presence 2002, Porto, Portugal, 2002

[24] P. Resnick and H.R. Varian, "Recommender Systems", Communications of the ACM, Vol. 40., No. 3, 1997. [25] R. J. Riding and S. Rayner, Cognitive styles and learning strategies. Understanding style differences in learning and behavior, London: David Fulton Publishers, 1998 [26] D. Riecken, "Personalized views on personalization", Communications of the ACM, vol. 43, 8, 27-28. 2000. [27] T. Saari, Knowledge creation and the production of individual autonomy. How news influences subjective reality, Reports from the department of teacher education in Tampere University, 1998. [28] T. Saari, Mind-Based Media and Communications Technologies. How the Form of Information Influences Felt Meaning, Acta Universitatis Tamperensis 834. Tampere University Press, Tampere, 2001. [29] T. Saari, Designing Mind-Based Media and Communications Technologies, Proceedings of Presence 2002 Conference, Porto, Portugal, 2002. [30] R. Schank and R.P. Abelson, Knowledge and memory: The real story, Advances in Social Cognition, Volume VIII, Lawrence Erlbaum Associates, Hillsdale, New Jersey, USA, 1995. [31] M. Turpeinen, Customizing news content for individuals and communities, Acta Polytechnica Scandinavica. Mathematics and computing series no. 103. Helsinki University of Technology, Espoo, 2000. [32] L. Weitzman and K. Wittenburg, "Automatic presentation of multimedia documents using relational grammars", Proceedings of Second ACM International Conference on Multimedia, San Francisco, 1994, pp. 443451.

[23] B. Reeves and C. Nass, The media equation. How people treat computers, television and new media like real people and places, Cambridge University Press, CSLI, Stanford, 1996.

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

10