A Personalized Emotional Intelligent Tutoring System Based on AI Planning Dissertation zur Erlangung des Doktorgrades Doktor der Naturwissenschaften (Dr. rer. nat.) der Fakulta¨t fu¨r Ingenieurwissenschaften und Informatik der Universita¨t Ulm vorgelegt von Heba Mohamed Atef Elbeh aus Menoufyia, A¨ gypten Institut fu¨r Ku¨nstliche Intelligenz Fakulta¨t fu¨r Ingenieurwissenschaften und Informatik Universita¨t Ulm Institutsleitung Prof. Dr. Susanne Biundo-Stephan Ulm, Deutschland June 2012 Amtierender Dekan der Fakulta¨t fu¨r Ingenieurwissenschaften und Informatik: Prof. Dr.-Ing. Klaus Dietmayer Vorsitzender des Promotionsausschusses: Prof. Dr. Uwe Scho¨ning Mitglieder des Promotionsausschusses: Prof. Dr. Michael Weber Prof. Dr. Wolfgang Minker Die Gutachter der Dissertation: Prof. Dr. Susanne Biundo-Stephan Prof. Dr. Harald Traue Tag der Promotion: 27 September 2012 A Personalized Emotional Intelligent Tutoring System Based on AI Planning A PhD thesis submitted to the Faculty of Engineering and Computer Science, Ulm University in fulfillment of the requirements for the degree of Doktor rerum naturarum (Dr. rer. nat.) by Heba Mohamed Atef Elbeh from Menoufyia, Egypt Institute of Artificial Intelligence Faculty of Engineering and Computer Science Ulm University Advisor Prof. Dr. Susanne Biundo-Stephan Ulm, Germany June 2012 Dean of the Faculty of Engineering and Computer Science: Prof. Dr.-Ing. Klaus Dietmayer Chairman of the doctoral committee: Prof. Dr. Uwe Scho¨ning Members of the doctoral committee: Prof. Dr. Michael Weber Prof. Dr. Wolfgang Minker Reviewers of the dissertation: Prof. Dr. Susanne Biundo-Stephan Prof. Dr. Harald Traue Day of Conferral of Doctorate: 27 September 2012 ABSTRACT The Intelligent Tutoring System(ITS) is a computer based learning system which assists stu- dents in their learning process. It has the ability to be adaptable according to the needs of students. In ITS, it is equally important to consider not only the cognitive level of the student but also the e- motional state of the student. Psychological researches indicate that emotions have a deep influence on the efficiency of memory storing and retrieving processes. Also, it is important to individualize the learning process and to teach students according to their personality types. Thus, students with various personality types differ in their cognitive capabilities, emotional and motivational states, learning style, appraisal, and coping way with incoming events. Although several approaches have been constructed for ITS, enhancing students emotional intel- ligence has not been considered so far. Emotional intelligence (EI) is considered as an important factor for increasing the student’s performance. The researches show that IQ (intelligence Quo- tient) contributes only about 20% to success in life, the rest of 80% of success depends on the EQ (Emotional Quotient). Thus, in our educational systems, we are not taught how to handle frustra- tion, anxieties, stress, and failure or depression problems during the learning process. EI is defined as the ability to recognize the meanings of emotion and their relationships, to reason about emo- tions, to enhance thought. With the consideration of previous factors, we present a novel framework for modeling an indepen- dent authoring/course generation system named PANDA.TUTOR. This system allows the author to prepare the course structure and content enriched with classification information without need to define the adaptation rules or specify configurations for each student, aiming to keep all the authoring effort low in terms of time and effort. The course generation system, which is basically a hierarchical planning system, generates a personal course for the student. The system enriches with different learning scenarios, teaching strategies and learning styles. The student’s personality type in our system serves as a guide for selecting an appropriate scenario, a teaching strategy, and a learning style in order to regulate with current emotional and motivational states of student. For that purpose, we developed general ontologies for representing student module and course module that can be used in different domains. These ontologies help the planning system to generate a personalized course. In addition, in order to enhance the student’s emotional intelligence, we de- veloped a dialog based HTN planning system. It is shown to improve the emotional intelligence of the student. This can be done by grasping the student’s appraisal to build up an individual’s interpretation of how external events relate to the student goal and desires, as well as overcoming their emotional reasoning. Emotional reasoning is defined as a thinking error that occurs when a student believes that his/her feeling is true regardless of the evidence, which prevents one looking at alternative, more balanced information or evidence. Different styles of dialog can be generated according to the student’s personality type. Moreover, we regard the student’s emotional appraisal and coping to regulate his/her emotion during the dialog and course generation. Thus, the objective of our approach is to teach the student during the learning process how to adjust or maintain the emotion, motivation, and cognitive states. Therefore, the proposed system is able to understand the student’s emotion and the reason behind this emotion to shape the students’ behavior. ACKNOWLEDGMENTS First of all, I would like to express my special appreciation and thanks to my advisor Prof. Dr. Susanne Biundo-Stephan, you have been a tremendous mentor for me. I would like to thank you for your continuous support and for all your invaluable guidance that makes this period of study enjoyable. Regardless of how much and what she had to do, she was always available for discussion and support, and a never dwindling source of ideas and suggestions. Beside my supervisor, I would like to thank the rest of my defense committee for their guidance. Special thanks to Dr. Bernd Schattenberg for his valuable support. I am grateful for the fruitful discussions with our friends in the Artificial Intelligence department, whose blessings and wishes have helped me come this far. Words cannot express the profoundness of my gratitude to my husband Dr. Mohamed Elkawkagy, my son Yousuf and my daughter Salma. Thank you for your support and patience during the whole time of my thesis. Without them, I wouldn’t be standing here. My sincere thanks go to my parents, sisters, for supporting me in everything, and especially I can’t thank you enough for encouraging me throughout this experience. Finally I thank my God, for letting me through all the difficulties. I have experienced Your guidance day by day. You are the one who let me finish my degree. Thank you, Allah. Ulm, Germany, 2012 Heba Elbeh DEDICATION To my parents, my husband and my kids (Yousuf and Salma).