Education & AI High level discussion points relating ‘techniques / tools’ to possible ‘projects’
Reinforcement Learning & Sequential Decision Making Choose sequence of questions, tasks or modes to optimize an individual's learning progress Provide encouraging prompts or reminders at beneficial times to optimize engagement Frictionless assessment, constantly monitor what a student answers correctly or incorrectly and realtime update an estimate of their knowledge.
Computer Vision & Speech Recognition Assessing students' concentration / frustration / stress-levels via facial expressions Speech recognition and synthesis to coach public speaking and teach languages
Generative Models – Bayes / Markov etc. Produce many novel exercises from few examples or rough templates Generate encouraging feedback messages for students Automatically produce quizzes from a knowledge base (e.g. Wikipedia Page → Quiz) New creative tools for art and literature classes
Natural Language Processing Conversational interfaces for interactive tutoring (Siri for Students) Automated marking of short answer questions or essays Answering students' free-form questions
Recommendation Systems & Topic Modelling Recommend course material based on similar or successful students' choices Automatically recommend revision material based on incorrectly answered questions Realtime rank students' aptitude Adaptively rank and match questions and homework by difficulty Adaptively categorize material by topic and style Match students and material by style and need Match students to peer learning groups by style and need Match students with teachers or tutors by style and need
Clustering / Exploratory Analysis Group students by how they learn Identify early-warning signs of under-performance Understand how schools, classrooms and teachers relate to each other Analyse course material for good structure and coverage of topics
No really AI But definitely still in scope
Analytics & Reporting Realtime reports on student, classroom and school performance Tracking of positive / negative indicators over time
Data Visualisation Explorable visualisations of student journeys Present student changes over time Present interactions with different teachers and topics.
Operations Research Optimal scheduling for classrooms, teachers, sports equipment or other resources.
Personalised Learning Some final discussion points
Learner Profiles Use clustering to identify groups of similar students and hidden patterns Make recommendations for intervention based on past evidence Alert on early-warning signs or large changes
Personal Learning Recommend follow-on material based on past interest and performance (e.g. Netflix recommendations for homework)
Competency Based Progression Interactive tutoring (e.g. Siri for Students) Question answering and Search (e.g. Google your textbook) Facilitate peer learning by automatically matching students with peers at similar (or different) levels Frictionless assessment, constantly monitor what a student answers correctly or incorrectly and real-time update an estimate of their knowledge.
Flexible learning environments Automatically and dynamically schedule and allocate resources based on recent data