(ALGS) ICELW 2020 June 10 th -12 th , New York, NY, USA AUTHORS - PowerPoint PPT Presentation

BALANCING THE ROLE OF MACHINE LEARNING AND TEACHER IN ADAPTIVE LEARNING GUIDANCE SYSTEM (ALGS) ICELW 2020 June 10 th -12 th , New York, NY, USA AUTHORS Ghada El-Hadad, Doaa Shawky and Ashraf Badawi Zewail City of Science and Technology/Center for Learning Technologies, Giza, Egypt

AGENDA ■ Problem Statement ■ The Two Main Pillars That ALGS Rests Upon ■ Why Is The Teacher’s Role Crucial in Adaptive Learning Systems? ■ Why Is The CSCL Important in Adaptive Learning Systems? ■ The Role of ML in ALGS ■ The Role of The Teacher in ALGS

Problem Statement ▪ Adaptive learning environments can have different flavors; some are intelligent tutoring systems, and some others are learning analytics ▪ Relying solely on deep learning came along with some concerns including motivation, procrastination, engagement and others. ▪ AI can perform tasks beyond human capabilities, which may result in complexity as the system would generate adaptation suggestions that the teacher may not be able to interpret. ▪ The novelty that ALGS is proposing is balancing the machine learning and the human factor in an attempt to reduce this gap.

The Two Main Pillars That ALGS Rests Upon ▪ ALGS is based on the teacher and the computer-supported collaborative learning (CSCL). ▪ Human guidance, and computer supported collaborative learning are suggested to work side by side with AI

Ad Adapti ptive e Learning ng Guidan ance ce System em ALGS DISCUSSION FORUMS LEARNIN NING G DECISI SIONS ONS Stude udent nt CSCL CL Teache cher GROUP TASKS REVIEW EW WORK CONTEN ENT RECOMM OMMEND ENDAT MODEL IONS Stude udent nt Stude udent nt

Why Is The Teacher’s Role Crucial In Adaptive Learning Systems? ▪ ALGS deploys the mentor’s physical existence with technology to maintain a healthy learning environment that corresponds to individual learners’ needs. Motivation, procrastination, engagement, and keeping cohesive learning ▪ environments online are noted as issues that can be better enhanced by personal relationships.

Why Is The CSCL Important In Adaptive Learning Systems? CSCL contexts: ▪ Facilitate group interaction among learners. Enable learners to exchange ideas and help each other to understand the ▪ topics and answer the test questions in a collaborative way.

The Role of The Teacher The Role of ML in ALGS In ALGS • Creating the database upon which the system • Building a reliable User Model, which is the key filtering function is based in the adaptation process, by collecting data • Feeding the system with data about learners about students’ actions and behavior patterns, that result from face-to-face interaction in the and then analyzing these large datasets. classroom • Detecting any deviation in the behavior of each • Suggesting an initial path since there are no learner from peer groups (CSCL feedback) and usage data stored yet in the system to adapt to alarming the teacher to intervene accordingly. particularly at the very early stages in ALGS • Generating automated recommendations to • Receiving system alarm and tracking the history students of the content they should study next of a particular student to interpret the rationale based on analyzing students’ learning patterns behind such deviation or failure to cope with the and behaviors. peers • Processing highly time-consuming analyses that • Reviewing and refining the system-generated are beyond human capabilities. recommendations • Tailoring adaptive recommendations to • Intervening and taking the proper action either individual learners’ needs based on the User in class or feeding adaptive decisions to the Model. system

Adapti Ad ptive e Learning ng Guidan ance ce System em ALGS Mach chine ne Teach cher er STEP EP 2 STEP EP 1 • Creates content nt model • Buil ilds ds user r model (from m psychome hometrics trics, CF, and CB Filteri ering) ng) (what at studen ents ts should d be learning) ning) STEP EP 3 STEP EP 4 • Reviews ws system m re recommenda ndations tions • Ge Genera rates s re recommendatio mendations ns • Makes learning ning decisio sions ns (based on student’s interaction)

REFERENCES [1] G. Weber, "Adaptive Learning Systems", Encyclopedia of the [8] R. Oliver, "Exploring strategies for online teaching and learning", Distance Sciences of Learning, pp. 113-115, 2012. Available: 10.1007/978-1-4419- Education, vol. 20, no. 2, pp. 240-254, 1999. Available: 1428-6_534 [Accessed 16 December 2019]. 10.1080/0158791990200205. [2] G. Magoulas, K. Papanikolaou and M. Grigoriadou, "A [9] Z. Mohamed, M. El Halaby, T. Said, D. Shawky and A. Badawi, connectionist approach for adaptive lesson presentation in a distance "Characterizing Focused Attention and Working Memory Using EEG", Sensors, vol. 18, learning course.", in International Joint Conference on Neural Networks. no. 11, p. 3743, 2018. Available: 10.3390/s18113743. Proceedings (Cat. No. 99CH36339), 1999, pp. (Vol. 5, pp. 3522-3526). [10] W. Shelley, S. Denise, N. Evangeline, O. Ruth and V. Evangeline, "From IEEE. Virtual Strangers to a Cohesive Learning Community: The Evolution of Online Group [3] [4]R. BAKER, "Challenges for the Future of Educational Data Development in a Professional Development Course", Journal of Technology and Mining: The Baker Learning Analytics Prizes", in 9th International Teacher Education, vol. 14, 2006. [Accessed 16 December 2019]. Conference on Learning Analytics and Knowledge, 2019. [11] B. Holmberg, A theory of teaching-learning conversation. London, UK: [4] D. Dżega and W. Pietruszkiewicz, "Intelligent Decision-Making Taylor & Frances, 2007, pp. 69-76. Support within the E-Learning Process", Intelligent and Adaptive [12] G. Stahl, T. Koschmann and D. Suthers, "Computer-supported Educational-Learning Systems, pp. 497-521, 2013. Available: 10.1007/978- collaborative learning: An historical perspective", Cambridge handbook of the learning 3-642-30171-1_20. sciences, pp. 409-426, 2006. [5] L. Wong and C. Looi, "Swarm intelligence: new techniques for [13] K. Vanlehn et al., "The Andes physics tutoring system: five years of adaptive systems to provide learning support", Interactive Learning evaluations.", in Proceedings of ICAIE, 2005, pp. 678 – 685. Environments, vol. 20, no. 1, pp. 19-40, 2012. Available: [14] U. Faghihi, P. Fournier-viger and R. Nkambou, "A computational model for 10.1080/10494821003714681. causal learning in cognitive agents", Knowledge-Based Systems, vol. 30, pp. 48-56, [6] [4]L. Kirtman, Kirtman, L. (2009). Online versus in-class 2012. Available: 10.1016/j.knosys.2011.09.005. courses: An examination of differences in learning outcomes.. Issues in [15] [10]G. El-Hadad, D. Shawky and A. Badawi, "Adaptive Learning Guidance Teacher Education, 18(2), 103-116., 2009. System (ALGS)", in 10th International Conference on Learning Analytics and [7] B. Tuckman, "The effect of motivational scaffolding on Knowledge, Frankfurt am Main, Germany, 2019. procrastinators’ distance learning outcomes", Computers & Education, vol. 49, no. 2, pp. 414-422, 2007. Available: 10.1016/j.compedu.2005.10.002.