The platform is used to assess and monitor the children' progress and performance in meeting preset learning objectives without any manual intervention.
Towards an Ontology-driven Game-based Educational Platform with Automatic Student Monitoring Sarah Malaeb, Aiman Erbad, and Jihad M. AlJa’am Department of Computer Science and Engineering, Qatar University, Qatar Abstract—This paper presents a new educational platform for young children. It is based mainly on serious games and semantic web. The platform is used to assess and monitor the children’ progress and performance in meeting preset learning objectives without any manual intervention. The learning games are used to improve the children’s learning outcomes and keep them motivated. The platform monitoring features allow the teachers to focus on the children’ achievement of every learning objective and empower also the parents’ engagement in their children’s learning experience. In fact, they can follow up the children and know their weaknesses and strengths. A new ontology is proposed to map the programs curriculums and learning objectives with the flow-driven game worlds’ elements. The children’ performance is evaluated through the ontology using information extraction with an automated reasoning mechanism that is guided by a set of inference rules. The platform is implemented in a 3-tier architecture where mobile game applications are used. These games can query and update the ontology in real time through a web service by invoking data management, reasoning, monitoring and reporting operations using Apache Jena Ontology API. The platform can be used to dynamically generate the content of the games based on the children’ preferences and acquired knowledge. Keywords— educational games, semantic web, student monitoring, rule-based reasoning, ontology information extraction.
I.
INTRODUCTION
Recent studies showed that students can become more motivated to learn with game-based learning tools [12][7]. These interactive elements facilitate problem solving [12] and make learning new concepts easier and encourage the students to work harder at school and also at home [4]. Active learning using game-based model forms the most effective learning methods for students. Educational games provide interactive experiences that motivate and actively engage the students in the learning process. They learn and practice the right way to achieve the objectives by working toward a goal, choosing actions, experiencing the consequences of those actions along the way and making risk-free mistakes. This guarantees longterm retention of information [3] which help the students increase their exam scores while acquiring the needed skills appropriately. TABLE I
ELEARNING COMPETING TECHNOLOGIES FEATURES
9 9
9
9 9 9
9 9
9
Progress Reporting
9 9
9
9 9
Digital Assessment
9
Features 9 9 9 9 9 9 9 Elementary Curricula (Mathematics, Sciences)
9 9 9 9
Elementary Age
Competing Technologies IXL Learning / HMH The Learning Company 9 Amplify Sylvan Learning 9 Knowledge Adventure® / Funbrain / 9 3P Learning / Arabic Rescue / ABCya.com / HMH Tribal Nova / EZSchool / Time4learning Moodle LMS / Atutor LMS DOKEOS LMS / eduTechnoz 9 The Magic School Bus 9 Rafed.Net / IslamiCity / Araboh / Sesame Street 9 Arabic 4 Fun / Baraem / Nick Jr. 9 Forma LMS / Hello-World Cricksoft Clicker Apps / Islamweb Rumie (Potential) Open edX LMS eFront LMS
Educational Games
Since the twenty-first century, digital technologies are increasingly supporting teaching and learning activities. Because learning is effective when it starts early, advanced early years’ educational tools are highly recommended to help new generations gain the necessary skills to successfully build opportunities and progress in life. With all the digital learning advances, there are still many problems that teachers, students and parents are facing. Students’ learning motivation, problem solving ability remain weak [12] while working memory capacity is found low for children under 11 years old which cause some learning difficulties [11], such as developmental coordination disorder, mathematics calculation, and language impairments. Another problem affecting the educational experience for young children is family engagement. Parents need to be more involved in the learning process and have quick and timely detailed feedback about their children’s progress in different topics of study. In fact, the schools days are limited and parents can play an important role in improving their children progress in learning and understanding concepts.
The traditional assessment tools provide global grading usually by topics of study (e.g., Algebra). Parents need a grading system by learning skills (e.g., addition facts to 10, solving missing-number problems, subtraction of money) to have a clear view about the specific skills that their children need to improve. Teachers need also an automated skills-based students monitoring tool to observe students’ progress in correspondence with the learning objectives to focus on personalized tutoring tactic and take accurate decisions. Such a tool allows the teachers to focus more on the students’ weaknesses and take the necessary actions to overcome with these problems.
9
We have conducted a survey and analyzed the features of 31 leading competing technologies in digital learning industry
978-1-4673-8243-4/15/$31.00 ©2015 IEEE 19-20 November 2015, Thessaloniki, Greece 2015 International Conference on Interactive Mobile Communication Technologies and Learning (IMCL) Page 141
as listed in Table I. We found that only 21 of them offer educational games, 22 are dedicated for elementary age range, 15 offer digital resources to support mathematics and sciences, 11 consider some digital tools for assessment to test children skills and 6 include automated progress reporting engine where most of them need manual data entry support from teachers. There is a need of a complete solution of a game-based learning platform with automatic performance and progress reporting without any manual intervention, and in particular customized to fit with elementary schools curriculum standards. In this work we present an educational platform called ‘IntelliFun’ that uses edutainment games to automatically monitor and asses the progress of elementary schools children. It can be applied to wider scope of outcome-based learning using games. The use of games brings direct advantages to children while the monitoring system brings support to parents and teachers to help improve children’s learning skills-based performance. An ontology-based approach is proposed to integrate curriculum standards and learning objectives with game worlds content (i.e., scenes and activities) to support students’ outcomes. This approach is known as Ontologydriven Educational Games with Automatic Student Monitoring OEGSM. It is also used to guide the process of information extraction based on a set of inference rules and predefined techniques to automatically evaluate and report on students’ progress. The remainder of the paper is organized as follows: Section II discusses some related works. Section III describes the platform implementation and finally section IV concludes the paper. II.
RELATED WORKS
Our platform integrates novel ideas that add value to the learning experience. OEGSM represents a fusion of many technologies: using ontologies in education, defining games worlds as flow-driven stories, using games in digital learning, using ontologies in game design, and integrating monitoring systems in education platforms. This section describes the technologies in OEGSM used to implement the IntelliFun platform and how it relates to the state of the art of some existing systems. A. Ontologies in Education [13] claims that ontology use in educational technology development started in 1999. Ontologies offer a wide range of benefits in education. They provide a link between the learning material and its individualized conceptualization results. They insure greater interoperability by securing learning objects reuse, information sharing and integration across different educational systems. They provide advanced rule-based reasoning querying tools for guidance and automation of any education related information extraction. Most of the ontologies created are in the field of education (31%) [1]. To design our curriculum-based ontology, we used a similar approach to the education ontology developed by the British Broadcasting Corporation (BBC) as the open source data model describing the United Kingdom national curriculum [5].
We applied the approach on the Qatar national curriculum using elementary mathematics as a case study. It is modeled at 3-dimensional space: fields of study, levels and topics. It has been designed to organize contents in a way that allows to navigate and discover learning resources and capture the sequential order of topics. In brief, the curriculum includes many fields of study (e.g. mathematics) which are taught in certain programs of study (e.g. mathematics for grade-1) which are defined by levels based on age groups (e.g. grade-1 for age range 6-7).The topics of study (e.g. reasoning and problem solving) include sub-topics (e.g., relationship between two entities) which are assigned to specific learning objectives (e.g., describing a simple relationship between two entities using appropriate mathematical terms). B. Student Monitoring in Learning Systems There are several research related work done in the field of students’ monitoring in learning systems which use different types of techniques to evaluate students’ progress such as latent semantic analysis, probabilistic and hierarchical latent analysis, semantic networks, fuzzy logic, procedural rules, and ontological inference reasoning. The learning assessment objects that are used include essays, text summaries, and unstructured written forum posts. A tool called STSIM, Semantic-web based Tool to Student Instruction Monitoring [2], uses student ontological network to monitor the student progress with respect to learning objectives during pedagogical activities execution. Students’ progress is measured by four key indicators: activity mark and weight, hits and errors. We followed the conceptual model of STSIM ontology to design our monitoring ontology using students’ traces to describe their progress during their learning experience in performing an activity. To evaluate students’ performance, we used the same key indicators with an inference rules-based reasoning engine to query correct, incorrect and incomplete actions performed by the students. C. Story-based Game Worlds and Education In the last few years, many methods defined game worlds as flow-driven stories in particular to algorithmically generate games contents. OEGSM analyzes games data as a set of islands where an island includes a set of scenes and every element is mapped with leaning contents and objectives. We used a method called Procedural Content Generation (PCG) [8]. Game worlds generation process in PCG uses the progression of the player through game space where game stories are represented as islands which are areas where critical actions occur and bridges which offer optional branch visit decisions. In our case, the system can hide the bridges which navigate to the islands assigned to the topics that their learning objectives have been already achieved by the students. For example, if the player has high level score in Algebra related to addition, only the bridges which are not linked to the islands of that topic will be accessible. Another related abstraction implements a course flow model as a tree hierarchy of story elements composed of scenes and actions [10].
978-1-4673-8243-4/15/$31.00 ©2015 IEEE 19-20 November 2015, Thessaloniki, Greece 2015 International Conference on Interactive Mobile Communication Technologies and Learning (IMCL) Page 142
D. Ontologies in Digital Games Advances in automated game design have given the possibility to develop game ontologies to create game architecture and add semantics knowledge to game objects, such as the avatars, the features of the avatars, and the places of interest [9]. Entika framework [6] generates game worlds at run-time while handling behavioral, physical and graphical semantics of game objects. OEGSM was built to enable dynamic generation of games contents. The game ontology that we have created follows the islands-based flow-driven model. III.
METHODOLOGY AND IMPLEMENTATION
In this section, we describe the design and implementation of our platform. We start by presenting our proposed OEGSM approach that we used to implement and develop our IntelliFun platform. Then we discuss the development of IntelliFun. A. OEGSM Approach OEGSM is a novel approach that uses ontology paradigm to incorporate learning and monitoring in serious games. Using ontologies in the field of education and games simplifies the challenging task of designing and linking educational and gaming data models based on predefined learning objectives and curriculum standards and providing advanced reasoning over the data. The OEGSM methodology comprises two modules: the ontology and the reasoning engine. It can be used in any education scope. We applied it to Qatari elementary schools. 1) OEGSM Ontology It is designed as a network of sub-ontologies related to the curriculum, the learning objectives, the game worlds and the student monitoring profile. It is illustrated in the class diagram (Figure 1). The curriculum and learning objectives ontology
Figure 1 OEGSM Ontology - Class Diagram
represents educational data including fields and programs of study where is taught a list of topics which have predefined learning objectives. The instances related to learning objectives are extracted from the elementary curriculum standards of the Qatar Supreme Education Council. The games ontology follows the technique of story-based scenario data model. A game is a set of islands where every island is assigned to a main topic of study (e.g. numbers and algebra) of a certain program (e.g. Mathematics for Grade-1). An island is a sequence of scenes where every scene is related to a specific sub-topic (e.g., addition facts to 10). The monitoring ontology stores the information tracked when a student is playing a game. This information include all the data which is measured to evaluate the students’ performance. 2) Assessment Reasoning Engine In the monitoring and assessment process, students’ performance is measured with an advanced reasoning engine which extracts the required information from data models. OEGSM solves difficulties faced in providing automated monitoring system delivering detailed outputs without any manual intervention. Ontologies are much known in their reasoning capabilities, and certain inferences rules should be defined at first in order to process this reasoning over the educational and gaming data. Many key indicators can be used to feed the reasoning engine. Data mining techniques can be used to generate students’ performance scores. While a student is playing a scene, a trace is saved in his/her monitoring profile which stores all the information indicating his/her performance (e.g., number of wrong actions, time spent, achievement status, etc.). Once the student completes a scene, the learning objective assigned with the topic of this scene is achieved. The reasoning engine evaluates the students’ performance by executing a rule of calculation that is pre-defined. IntelliFun currently follows a simple evaluation rule which takes into consideration two measurement attributes. It compares the time spent with the average time allowed to achieve the scene. Then it adds the number of wrong actions done before entering the correct answer. A scale matrix is defined for grading where the grades decrease with the increase of the wrong actions and the time spent. The reasoning engine will be extended to include more complex data mining techniques for evaluation. In the next step, we show how to select the most appropriate techniques for mining educational data to analyze students’ performance. These techniques can include more complex key indicators such as semantic indicators (e.g., knowledge, comprehension: is the student has low grade because of some misunderstanding of the question or of missing knowledge). B. IntelliFun Implementation IntelliFun is a game-based educational platform offering customized games to the students and skills-based automated monitoring system to their teachers and parents. Having these features integrated in one technology makes IntelliFun a creative solution in digital education. IntelliFun is deployed using 3-tier architecture (Figure 2) in order to assure reusability, dynamism and interoperability. The three main components developed are the ontology at the data layer, a web service at the business logic layer and the games at the presentation layer. The OEGSM ontology has been created
978-1-4673-8243-4/15/$31.00 ©2015 IEEE 19-20 November 2015, Thessaloniki, Greece 2015 International Conference on Interactive Mobile Communication Technologies and Learning (IMCL) Page 143
Figure 2 IntelliFun Platform - Deployment Diagram
with Protégé tool. Client mobile game applications are developed using Oracle Mobile Application Framework to run on any device including Android and iOS. Different users can access the same information from different mobile devices such as the teachers and the parents of the same student. For this reason, the ontology and the assessment component should be located on a remote server. The mobile apps access the remote ontology through SOAP web service which uses Jena Ontology API. It executes ontology management, monitoring and reporting operations. The main actors of IntelliFun are students, teachers and parents. The use case diagram in Figure 3 describes the main roles of the actors and the system. The students play the games related to the programs they are enrolled in. The system dynamically generates for them at run-time the content of games. The assessment component generates skills-based performance reports. The parents observe their children’s progress to know their weaknesses and strengths. The teachers monitor their students’ performance in correspondence with the learning objectives to allow them to take appropriate decisions to improve their skills as needed. The presentation diagram in Figure 4 presents the navigation between the screens of the mobile app.
Figure 3 Use Case Diagram
Figure34IntelliFun Presentation and Navigation Diagram Figure Platform - Deployment Diagram
978-1-4673-8243-4/15/$31.00 ©2015 IEEE 19-20 November 2015, Thessaloniki, Greece 2015 International Conference on Interactive Mobile Communication Technologies and Learning (IMCL) Page 144
IV.
CONCLUSION AND FUTURE WORK
We have presented a new educational platform that can be used to monitor and assess the students’ progress through edutainment games. It sustains educational objective and maintains the interactive fun experience. It focuses on learning objectives to improve the students’ learning outcomes. The ontology maps the game worlds’ content with the curriculum and learning skills. A rule-based reasoning engine reports detailed feedback about students' performance deducted from indicators gathered during the game playing experience. This platform will allow to perform research in the benefits of game-based learning for the elementary curriculum in Qatar. In future work, we will follow 4 major steps. To start testing our platform with real users, we will finalize the development of the first prototype of the mathematics curriculum case study. Then we will validate our proposed solution using a user study with students, parents and teachers who will answer an evaluating questionnaire after testing the technology. This will help us evaluate the efficacy of the platform and ascertain its benefits by analyzing its direct impacts in improving students' learning experience. Then we will extend the development of IntelliFun to consider two additional components: adding data mining techniques to the reasoning engine to define rules to evaluate students’ performance with complex performance key indicators, and considering dynamic generation of game worlds’ content based on students’ preferences and acquired learning skills. REFERENCES [1]
S. Dhuria, S. Chawla; “Ontologies for Personalized E-Learning in the Semantic Web”, IJAENT International Journal of Advanced Engineering and Nano Technology, Vol. 1, March 2014.
[2]
[3]
[4]
[5] [6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
H. Y. Corral, J. C. Parraga, “STSIM: Semantic-web Based Tool to Student Instruction Monitoring”, SCITEPRESS Science and Technology Publications, 2014, in press. C. J. Miller , M. J. Metz, “A comparison of professional-level faculty and student perceptions of active learning: its current use, effectiveness, and barriers”, Advances in Physiology Education, Vol. 38, September 2014. L. Kitching, S. Wheeler, “Playing Games: Do Game Consoles have a Positive Impact on Girls' Learning Outcomes and Motivation?”, European Journal of Open, Distance and eǦLearning, Vol. 16, October 2013. British Broadcasting Corporation, Curriculum Ontology, http://www.bbc.co.uk/ontologies/curriculum, October 2013. J. Kessing, T. Tutenel, R. Bidarra, “Designing Semantic Game Worlds”, in Proceedings of the 3rd workshop on Procedural Content Generation in Games, ACM New York, 2012. S. Kinash, J. Brand, T. Mathew, “Challenging mobile learning discourse through research: Student perceptions of Blackboard Mobile Learn and iPads”, Australasian Journal of Educational Technology, Vol. 28, May 2012. K. Hartsook, A. Zook, S. Das, M. O. Riedl, “Toward supporting stories with procedurally generated game worlds”, Conference on Computational Intelligence and Games (CIG), IEEE, p. 297–304, 2011. M. Roman, I. Sandu, S. C. Buraga, “OWL-Based Modeling of RPG Games”, Studia Universitatis Babes-Bolyai Series Informatica, vol. 56, 2011. L. Juracz, “Developing courses with HOLORENA, a framework for scenario- and game based e-learning environments”, International Journal of Software Engineering & Applications, Vol. 1, p. 1, October 2010. J. Holmes, S. Gathercole, D. Dunning, “Poor working memory: impact and interventions”, Advances in Child Development and Behaviour, Vol. 39, p. 1-43, January 2010. G. Cheng, “Using game making pedagogy to facilitate student learning of interactive multimedia”, Australasian Journal of Educational Technology, Vol. 25, p.204-220, May 2009. D. Dicheva, S. Sosnovsky, T. Gavrilova, P. Brusilovsky, “Ontologies and Semantic Web for E-Learning”, International Handbooks on Information Systems, Springer, pp 47-65, 2008.
978-1-4673-8243-4/15/$31.00 ©2015 IEEE 19-20 November 2015, Thessaloniki, Greece 2015 International Conference on Interactive Mobile Communication Technologies and Learning (IMCL) Page 145
