Learning Environments with Conversational Agents
Art Graesser, University of Memphis, United States
Implementing Learning Analytics
Barbara Wasson, University of Bergen, Norway
Intelligent Mentoring: Linking Learning Analytics and Interactive Nudges to Support Life-wide Learning
Vania Dimitrova, School of Computing, University of Leeds, United Kingdom
Learning Environments with Conversational Agents
Art Graesser
University of Memphis
United States
Brief Bio
Art Graesser is professor emeritus in the Department of Psychology and the Institute of Intelligent Systems at the University of Memphis, and Honorary Research Fellow at University of Oxford. His research is in discourse processing, cognitive science, and education. He has developed software in learning, language, and discourse technologies, including systems that hold a conversation in natural language with computer agents (AutoTutor) and that analyze text on multiple levels of language and discourse (Coh-Metrix). He served as editor of Discourse Processes and Journal of Educational Psychology, as president of Society for Text and Discourse and International Society for Artificial Intelligence in Education, and on four panels with the National Academy of Sciences and four OECD expert panels on problem solving (PIAAC 2011, 2021; PISA 2012, 2015).
Abstract
Computer agents can participate in conversations with students, help them learn, and assess their progress in acquiring knowledge and skills. The design of an agent (e.g., talking head, chatbot, avatar) requires a research team with expertise in artificial intelligence, computational linguistics, psychology, discourse processes, and data mining in addition to education. This presentation describes some learning environments with computer agents that target different adult populations. These systems implement the eight learning affordances in the digital age that were identified in the 2018 edition of How People Learn of the National Academy of Sciences, Engineering, and Medicine: interactivity, adaptivity, feedback, choice, nonlinear access, linked representations, open-ended learner input, and communication with others. Some systems have three-party conversations, called trialogues, where two agents (such as a tutor and a peer) interact with the adult. The adult populations have ranged from struggling adult readers who need to improve their comprehension skills to advance their careers (AutoTutor-ARC) to high ability sailors in the Navy learn electronics (ElectronixTutor). This presentation addresses a fundamental question: When does it make sense (or NOT) to include conversational agents in a learning environment?
Implementing Learning Analytics
Barbara Wasson
University of Bergen
Norway
Brief Bio
Professor Dr. Barbara Wasson is Director of the Centre for The Science of Learning and Technology (SLATE), the Norwegian national centre for learning analytics. She was one of the founders of Kaleidoscope, a European Network of Excellence on Technology Enhanced Learning. Wasson is/has been PI for numerous national and international projects, including the Erasmus+ project DALI: Data Literacy for Citizens. Currently she is a member of the Norwegian Ministry of Education’s Expert group on Learning Analytics, the Council of Europe Expert Group on AI in Education, and Norway’s representative to the European School Net’s Data Interest Group. Her research interests include learning analytics, AI and education, learning games, e-assessment, teacher inquiry, and data literacy.
Abstract
Implementing Learning Analytics in Education
After 12 years Learning Analytics and Knowledge is a thriving research field, but still has to make a major impact on education. Why is it so hard to implement learning analytics in education? Through reflections on research to practice, and examples from my own experience in Norway I will address the complex challenges (e.g., law, ethics, technological, cultural, and pedagogical) that are met when scaling up the use of learning analytics in education.
Intelligent Mentoring: Linking Learning Analytics and Interactive Nudges to Support Life-wide Learning
Vania Dimitrova
School of Computing, University of Leeds
United Kingdom
![]()
Brief Bio
Vania Dimitrova is a Full Professor (Chair) of Human-Centred Artificial Intelligence at the School of Computing, University of Leeds, UK. Her research focuses on building systems that help people make sense of data, take decisions in complex settings, expand their knowledge, and learn from experience. She is currently President of the International AI in Education Society, Co-Director of the UKRI Centre for Doctoral Training in AI for Medical Diagnosis and Care, and member of the the Council of Europe Expert Group on AI in Education. She is Associate Editor of both the International Journal of AI in Education, and Frontiers of AI: AI for Human Learning and Behavior Change. She was Associate Editor of IEEE Transactions on Learning Technologies (IEEE-TLT), a member of the editorial boards for the personalisation journal (UMUAI), and chaired several international conferences in intelligent learning environments (AIED, ECTEL, ICCE).
Abstract
In the past decades, the successful and deployable solutions of intelligent learning environments have predominantly been in formal education, usually school or university settings. Socio-economic and technological developments show the need to broaden the educational scope to address soft (transferable) skills and to include life-wide learning in real contexts and everyday life experiences. I will argue that this calls for a change in the way we design intelligent learning environments, shifting from tutoring to coaching and mentoring. Digital transformation of education enables the capturing of digital traces which can be analysed to understand the learners, their behaviour, and the overall learning process. Hence, learning analytics approaches can provide ‘sensors’ about aspects of the learners and the learning context. This can then indicate situations in which we can provide nudges to foster behaviour change and facilitate learning. I will illustrate this in several domains and will further research directions.