Agathe Merceron is Professor of Computer Science at the Beuth University of Applied Sciences Berlin. She is the head of the Media Informatics Online Bachelor's and Master's degree programs. Her research interests lie in Technology Enhanced Learning with a focus on Educational Data Mining and Learning Analytics. She is nationally and internationally involved in these areas and has served as program-chair for both the international conference on Educational Data Mining (EDM) and the international conference of Learning Analytics and Knowledge (LAK); she is Associate Editor of the Journal of Educational Data Mining.

Keynote : Imperfection in Learning Analytics

Learning analytics is defined as “the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimising learning and the environments in which it occurs”. However, the results of our analyses are not perfect. Prediction of marks or of drop-out for example, has almost never an accuracy of 100%. Kitto, Buckingham Shaw and Gibson ask an important question: “It is usually assumed to be important that classifiers be accurate, as otherwise a student will be subjected to inappropriate interventions. However, such a position leaves us in a dilemma; are we to wait until perfect accuracy is achieved?” (LAK 2018). In this talk, I will present research work that is confronted with this dilemma.

Han van der Maas (1966) studied Psychology at the University of Amsterdam and received his Ph.D in 1993 for research on methods for the analysis of phase transitions in cognitive development. In 2005 he became professor and chair of Psychological Methods group at the University of Amsterdam. In 2009 he founded Oefenweb.nl, a spin-off company of the UvA-holding, selling a unique online adaptive learning and monitoring game platform for children. The data has been used for numerous publications and several dissertations. His research focuses on formalizing and testing psychological theories in areas such as cognition, expertise, development, attitudes and intelligence. His 2006 paper on the mutualism model of general intelligence was the start of the network psychometric approach that has become very popular in different areas of psychology.

Keynote : Learning analytics in online adaptive training and monitoring systems

We developed Math Garden in order to collect highly frequent and reliable data on children’s cognitive development. Participants do their daily practice in our online adaptive learning environment, meaning that they always do exercises adapted to their own ability level. This accurate adaptability was achieved by developing a novel adaptive algorithm based on the Elo rating system for chess, where item difficulties and student abilities are estimated on the fly.There are numerous benefits of this approach. First, it is relatively cheap to build such systems as item difficulties do not have to be pretested. Second, our learning analytics are based on the highest quality measurement scales available in educational measurement. Third, we create a win-win-win setup. Children like to play our games and practice relevant materials, teachers do not have to grade exercises and are provided with detailed learning analytics, and researchers get access to a rich database on cognitive development. Math Garden is used by about 200,000 children in the Netherlands. We collect about 1 million item responses per day, and we published over 30 peer-reviewed papers using Math Garden data. In this talk I will introduce the Math Garden approach, present results, and discuss new developments such as randomized double blind experiments and new learning analytics based on mouse tracking.

Claudio Cimelli is head of the digital projects incubation mission (Numéri'lab) since 2014. Previously, he was coordinator of the national digital dialogue (from November 2014 to May 2015). He pilots programs and research / study actions of the Directorate of Digital Education (French ministry of education) :

• - 10 national reflection programs including data analysis, adaptive learning, collaborative work and digital immersion.
• - Management of partnerships with companies and web publishers.
• - Supervision of the DNE experimentation programs, dealing with new technologies such as Big data and AI.

Vice-Rector of CNED, Jérôme Villot has stared his teaching career in 1999. From 2006 to 2015 he led the management of several ESCEM’scourses and programs. He has developed and deployed innovative pedagogical methods, in particular the adaptive method of anticipatory pedagogy introduced in 2008 on the Bachelor program. In 2014, he became ESCEM’s Director of Programs and Education. At the same time, he continued to carry out teaching activities, especially in economics and mathematics. Since 2004, he has also been involved in consulting in business creation and development. Jérôme Villot was born in 1972. As deputy director at the CNED since 2016, Jérôme Villot has led and implemented the school's core tasks : distance learning, massively, with a focus on the success of learners. The aim is to improve, professionalize and evolve this activity to adapt it to a profoundly changing context.

In parallel with a doctorate with Prof. Guy Pujolle, Julien participated, as responsible for research and development, then assistant technical director, to the creation and development of the startup UCOPIA, become leader in France and Europe in its field. Julien is currently Technical Director (CTO) of Nomad Education, a free mobile application that assists middle school students and students towards success and helps them become actors in their future. He is also in charge of new projects such as Adaptive Learning, supported by the BPI.

Graduate of ESSEC in 2007, Clémence began her career in digital marketing at Lagardère and joined the world of start-ups. After 10 years of experience in the areas of e-commerce, sustainable development and health, she wants to engage in a project with real social impact. She joined Nomad Education in 2018, as Marketing and Communication Director, to help grow this Education project.

Ivanna is a 3rd year student at the University of Amsterdam, Netherlands. She is heavily involved with university life, as a student assistant for one of the courses and was part of the student council. She is interested in education and possibilities that modern technologies can offer in this field.

Azim Roussanaly is an associate professor in computer science at Université de Lorraine. His research activities focus on data, personal data and privacy, when storing and managing data.

Thomas Toulotte is a PhD student in Law at Université Versailles Saint Quentin. He is specialised in digital law in the frame of Learning Analytics.