Towards a Generic Socio-cultural Profile for Collaborative Environments

1 downloads 0 Views 228KB Size Report
This paper focuses on socio-cultural user modeling in a social and intercultural collaborative context. we first, study and discuss several examples of user.
2011 IEEE International Conference on Privacy, Security, Risk, and Trust, and IEEE International Conference on Social Computing

Towards a generic socio-cultural profile for collaborative environments Ouamani fadoua

Narjès Bellamine Ben Saoud

RIADI Laboratory ENSI

RIADI Laboratory ENSI

Manouba, Tunisia [email protected]

Manouba, Tunisia [email protected]

Henda Hajjami Ben ghezala RIADI Laboratory ENSI

Manouba, Tunisia [email protected] using a user modeling approach that will allow us to build a generic and multidimensional socio-cultural user profile that show how socio-cultural factors can affect all aspects of the human behavior, human personality, human perception, etc. The construction, the implementation and the validation of this model in our future work will allow us to personalize and adapt collaborative environments to the user, taking into account his social and cultural background.

Abstract This paper focuses on socio-cultural user modeling in a social and intercultural collaborative context. we first, study and discuss several examples of user models in many research domains; we mainly focus on how these models are defined and represented. After that, we describe our generic and multidimensional socio-cultural profile, so, we define and justify its dimensions, its characteristics and its properties. We also describe the way the data to fill the model will be collected and updated. Finally, we discuss what we have presented and we give a brief plan of our future work to better address these important issues.

In this paper, we first, study and discuss several examples of user models in many research domains; we mainly focus on how these models are defined and represented. After that, we describe our generic and multidimensional socio-cultural profile, so, we define and justify its dimensions, its characteristics and its properties. We also describe the way the data to fill the model will be collected and updated. Finally, we discuss what we have presented and we give a brief plan of our future work to better address these important issues.

Key words user modeling, socio-cultural factors, user profile, collaborative environments

I.

INTRODUCTION

With the expansion and the huge use of the internet and information technologies, collaborative environments have sprung in many domains like learning, work and game. The goal of such environments is to support the collaboration between two or more people in order to make effective the interaction between them and to improve its outcomes. However, “the collaboration is a social structure in which two or more people interact with each other and in some circumstances, some types of interaction occur that have a positive effect”[1]. Besides, these environments bring together and link people from around the world. So, they are from different cultures and social contexts. They don’t have the same needs, the same behavior, the same interests, the same goals, etc. Therefore, they aren’t waiting for the same things from the system through which they interact together. We have to take into account these socio-cultural characteristics of both people and the interaction between them while designing such environments. In our thesis work, we choose to address these issues

978-0-7695-4578-3/11 $26.00 © 2011 IEEE DOI

II.

STUDY OF SEVERAL USER MODELS IN MANY DOMAINS

Kobsa [2] argued that user modeling is usually traced back to the works of Allen, Perrault and Cohen [3, 4, 5] and Rich [6] (mentioned in [2]). Since that time, many researches were inspired by their seminal works to develop many systems that collected different kind of data about the current user, and adapted to the user in different ways. User profiling has been the subject of many user modelins studies. It is a technique to represent user model by a set of characteristics, these characteristics can be related in different manners. It turns out that it’s a very complex problem. This evidence was proved by information and human sciences like psychology, sociology, AI (Artificial Intelligence), HCI (HumanComputer Interaction), pedagogy, etc. In this section, we have selected several examples of user 599

profiles in many research areas. We have focused mainly on how researchers have extracted and defined characteristics of their profile and represented it.

extract the learning preference of the learner and a MCQ (Multiple Choice Question) to extract his/her knowledge level. D. Collaborative environments Specifically, in collaborative environments, the researchers use mostly stereotypes to model users. We have studied the AUTO-COLLEAGUE [11] system, which is an adaptive CSCL (Computer Supported Collaborative learning) system that builds user stereotypes to provide an adaptive and personalized environment for the learning of UML in groups. The system traces the performance of learners and makes inferences about user characteristics such as the type of their performance, their personality and the level of their expertise. The characteristics are defined based on cognitive and psychological theories.

A. IR (Information retrieval) In the IR research domain, user modeling is a recurrent issue. It consists mainly in the integration of the user profile in the IR process to better meet the user information needs. The majority of user profiles offered represent the user interests, user preferences and his/her research purposes. In [7], the authors propose a user profile of researchers as an extension of the FOAF ontology [8]. Based on the analysis of researchers personal web pages, and on their DBLP bibliography they extract and define the main characteristics: researcher characteristics (fax, phone number, address, email, affiliation, position, photo, etc.), publications characteristics (title, date, editor, download URL, etc.), their interests and finally their social relationships and related type. Therefore, they build, for this situation, a profile of 4 concepts, 29 properties and 4 relations.

E. Discussion We have presented a series of recent researches that proposes user models in many domains. These researches have used different techniques to represent their models (profiles, stereotypes, etc.) and define its characteristics. In our study of user modeling literature, we have noticed the absence of the socio-cultural characteristics in the user models proposed. And even the few researches that deal with such issues are treating it partially with some characteristics in the model. So they ignore the very important notion of the influence of the sociocultural dynamic on our behavior, cognition, emotion, personality, motivation, etc which is important mainly in collaborative and social environments because it allow us to understand and handle the mental processes and the behavior of the human beings. This was demonstrated by Vigotsky [12] in his socio-cultural thesis. In the next session, we present and define the socio-cultural dynamic as a user profile with socio-cultural dimensions which explains the influences of socio-cultural factors on human being.

B. HCI Researches in HCI domain have been interested in user modeling to improve the collaborative nature of HCI systems. Here is a selected example from a recent work [9] in which the authors propose a perception model, a cognitive model and a motorbehavior model. The models are parameterized to represent different physical abilities, levels of skill and input devices. The perception model simulates the phenomena of visual perception (like focusing and shifting attention). It can also simulate the effects of different visual impairments (e.g.: wet and dry Macular Degeneration, Diabetic Retinopathy, Tunnel Vision, etc.) on interaction. The cognitive model simulates expert performance and performance of novices by using a dual-space model. The motor-behavior model is developed by statistical analysis of cursor traces from motorimpaired users. As part of the model, we also develop a new scale of characterizing the extent of disability of users by measuring their grip strength, which was not earlier possible by using clinical scales. These models do not need knowledge of psychology or programming to operate. They have graphical user interfaces to provide input parameters and showing output of simulation.

III.

SOCIO-CULTURAL CHARACTERISTICS OF OUR PROFILE

We describe in this section our multidimensional profile composed of four dimensions: cultural characteristics, social characteristics, individual characteristics and cognitive characteristics. Each dimension contains a number of characteristics and each characteristic is a group of attributes. To extract and define the dimensions, their characteristics and their attributes, we have studied many researches in human and social sciences (anthropology, ethnology, sociology, psychology) that show the cultural effects on human characteristics, behavior and mental and cognitive faculties.

C. E-learning Generally, the user modeling in e-learning models a specific user, the learner, in order to adapt the knowledge representation to his/her learning preferences and help him/her to have the better learning experience. In [10], the author proposes a learner profile composed of 4 concepts: the user identifications (first name, last name, language, etc.), his/her knowledge level, his/her learning preferences and his history of reading documents. The author uses for example, cognitive sciences to

Figure 1. A therotical framework for construction of a sociocultural user profile

600

C. Individual characteristics As we have cited above, individual has some characteristics that we cannot judge or predict based on behavior patterns [20]. According to Oishi [19], theses individual characteristics are personality, biological state, goals and preferences. These characteristics are also under cultural and social effects [19, 21]. Individuals built their self-concept, peer-concept, life philosophy, progressively by observing their behaviors, the other behaviors and the reaction of others to their behaviors. This mechanism forms their personalities. They also develop their own temperament and values [19]. [19] argues that biological state (drug, illness, disable) and individual temperament influence the emotional and cognitive state and behaviors of individuals, such an influence can be decreased or increased by social factors like parents, social situation, social role, etc. In [21], the author clarifies also the important of goals in social and cultural life. The achievement of goals influences our daily well-being and brings positive emotions and life satisfaction. The goals are an integral part of self-construction [22]. These studies allow us to distinguish individual characteristics which are personality, biological state, preferences and goals and some attributes of personality which are selfconcept, peer-concept, temperament, life philosophy and values.

A. Cultural characteristics The cultural characteristics are inherited from the cultural group to which user belongs. Such a profile was described by many anthropologists and ethnologists using national models of culture like the one proposed by Hofstede [13], the one proposed by Hall [14] and others proposed by Schwartz [15], Kluckhohn [16], the globe model [17], etc. these model presented the cultural characteristics using cultural dimensions or values. These dimensions or values are measured through questionnaires in order to calculate rates for each nation that measure the intensity of the presence of each dimension or value within the population of each nation. We have done a comparative study of the convergence and divergence of these several models of culture. As a result of this study, we have identified 5 common dimensions which are: the distribution of power and authority [13], the belonging (individualism/collectivism)[13], the relation with the environment, the attitude toward time [14] and the control [13].

D. Cognitive characteristics According to the authors in [23] and as we have seen above, there is a mechanism of mutual influences between characteristics: socio-cultural factors influence personality, mental state, cognitive state, goals, preferences, etc. The authors in [23] argue that the emotional and motivational [24] [25] state influence our way of reflection and reasoning about a thing. According to these studies, we can define the cognitive characteristics as follow: we have the emotional state, the motivational state and the way of reasoning or reflection.

B. Social characteristics In [18], the author argues that the human being is both the result of the social environment, the surrounding culture and a progressive self construction. The interaction between individuals, the society and the culture define behavior patterns, even of their deepest emotions. The author of [19] explains that individual has a social situation and interacts with his society by doing his role in it. These two social characteristics don’t affect individuals uniformly because there are individuals who like their role and they seek to achieve it and others dislike it and avoid it. The culture gives sense to social roles and situations [18, 19]. Within a society, we can find patterns of social behavior but each individual has an individuality that distinguish him/her from other individuals [18, 19], so he/her has hi his/her own personalities and his/her social needs. So if we try to summarize social characteristics according to these studies, we found three characteristics: role (genre, family role, professional role, etc.), situation (family situation, financial situation, marital status) and needs

E. Data acquisition to construct and updating the model As we have seen in section 2, data acquisition can be done explicitly or implicitly. We have chosen to combine the two techniques to collect data about the users. Explicitly, we will use questionnaire from the social, cultural and psychological researches discussed above. For instance to measure the cultural characteristics, we can use Hofstede scores for each nation. Besides, to collect data about the personality, we can use a personality questionnaire such as Myers Briggs Type Indicator [25]. For social characteristics, the users can fill forms integrated to the applications or we can interviewer them. Implicitly, we will use algorithm and learning techniques to observe, save and learn the real behavior of the users in different case, in short term and long term to fill the other characteristics like cognitive characteristics and to update and evolve

601

[11] K. Tourtoglou, M. Virvou, “Evaluating an Intelligent Collaborative Learning Environment for UML”, ICSOFT 2010, p, 462-467 [12] I. Ivic, “LEV.S Vigotsky”, revue trimestrielle d’éducation compare (Paris, UNESCO : Bureau international d’éducation), vol. XXIV, n° 3/4, 1994, p 91/92. [13] G. Hofstede, “Culture’s consequences: International differences in work-related values”, Beverly Hills, CA:Sage, 1980. [14] E. Hall, “Beyond culture”, Anchor Press, New York, 1976 [15] S. H. Schwartz, “Value dimension of culture and national difference”, retrieved at: http://www.uib.no/psyfa/isp/diversity/content/reseach/multi cultur/Workshop/Schwartz%20.pDf, 1992. [16] F .Trompenaars and C. Hampden-Turner, “The Seven Cultures of Capitalism: Value Systems for Creating Wealth in the United States”, Japan, Germany, France, Britain, Sweden, and the Netherlands, McGraw-Hil, 1993. [17] R.J. House, P.J. Hanges, M. Javidan, P. Dorfman and V. Gupta, “Culture, Leadership, and Organizations: The GLOBE Study of 62 Societies”, Sage Publications: Thousand Oaks, CA, 2004. [18] R. Linton, ”Le fondement culturel de la personnalité”, Un document produit en version numérique par Jean-Marie Tremblay, professeur de sociologie au Cégep de Chicoutimi, 1845. [19] S. Oishi, “Personality in culture : a neo allportian view”, Journal of Research in Personality, vol. 38, 2004, p 68-74 [20] G.W. Allport, “Pattern and growth in personality”, New York: Holt, Rinehart & Winston, 1961. [21] V. Benet-Martínez, “Cross-cultural personality research: Conceptual and methodological issues”, In R.W. Robins, R.C. Fraley, & R. Krueger (Eds.), Handbook of research methods in personality psychology. New York, NY: Guildford Press, 2007. [22] H. C. Triandis, “The influence of culture on cognitive processes”, Illinois Univ at Urbana Group Effectiveness Research Lab, 2005 [23] S. Oishi, and E. Diener, “Goals, culture, and subjective well-being”, Personality and Social Psychology Bulletin, 27, 2001, p 1674-1682. [24] A. Norenzayan and R.E. Nisbett, “Culture and causal cognition”, Current Directions in Psychological Science, 9, 2001, p 132-135. [25] I.B Myers, M.H. McCaulley, N.L. Quenk and A.L. Hammer, ” MBTI Manual (Auide to the development and use of the Myers Briggs type indicator). ConsultingPsychologists Press (3rd ed), 1998. [26] A. Kobsa, “Generic User Modeling Systems”, User Modeling and User-Adapted Interaction 11(1-2), 2001, p49-63.

the initial profile obtained with the implicitly manner. To update the model in a collaborative context, the system has to observe the behavior of the user when interacting with the system and his/her behavior when interacting with the other users through the system. IV.

DISCUSSION AND FUTURE WORK

Our first contribution is to propose a generic and multidimensional socio-cultural profile. Why and how it can be a generic model? According to Kobsa [30] definition of the genericity, a generic model is a domain independent model, so it can be usable in many tasks, in many applications, in many domains to provide many services. It’s also expressive, so it contains the biggest number of information kinds and it’s allow strong inference capabilities. Our model is domain independent and it’s expressive because it’s a multidimensional model with 4 dimensions that define the whole human mechanism and the relations of influences between the components of such a mechanism. So it does will allow us to make strong inferences. In our future work, we will do a complete state of the art of the field of user modeling, we will look also more closely at the construction approaches of user models, specifically ontology. Therefore, we can begin the step of model construction using ontology, define the inference rules and the updating techniques cited above in practice. REFERENCES [1]

P. Dillenbourg, D. Traum, and D. Schneider, “Grounding in multi-modal task-oriented collaboration”, presented at European Conference on AI in Education. Lisbon, Spetember, 1996. [2] A. Kobsa “Generic user modeling systems”, in P. Brusilovsky, A. Kobsa and W. Nejdl. “The adaptive web: Methods and strategies pf web personalization, Berlin, Heidelberg, New York: Springer Verlag, 2007, p 136-154. [3] J.F. Allen, “A Plan-Based Approach to Speech Act Recognition”, Dept. of Computer Science, University of Toronto, Canada, Technical Report 131/79, (1979). [4] P.R. Cohen and C.R. Perrault, “Elements of a Plan-Based Theory of Speech Acts”, Cognitive Science 3, 1979, p 177212, DOI 10.1016/S0364-0213(79)80006-3. [5] C.R. Perrault, J.F. Allen and P.R. Cohen, “Speech Acts as a Basis for Understanding Dialogue Coherence”, Department of Computer Science, University of Toronto, Canada, Report 78-5, 1978. [6] E. Rich, “User Modeling via Stereotypes”, Cognitive Science 3, 1979, p 329-354. [7] J. Tang, L. Yao, D. Zhang and J. Zhang, “A combination approach to web user profiling”, presented in ACM Transactions on Knowledge Discovery from Data, Vol. V, No. N, March 2010. [8] D. Brickley and L. Miller, “Foaf vocabulary specication”, In Namespace Document , 2004 [9] P. Biswas and P. Robinson, “A brief survey on user modelling in HCI”, presented in IHCI 2010. [10] A. Behaz and M. Djoudi, “Ontological MODELING FOR THE CREATION OF A adaptive hypermedia”, MODELISATION ONTOLOGIQUE POUR LA CREATION D'UN HYPERMEDIA ADAPTATIF, presented in ticemed, 2009

602