Adaptive Learning
Overview
Adaptive learning studies how models and instructional systems can change their behavior as they encounter new data, shifting environments, or learners with evolving knowledge. Educational systems use learner-state models such as Bayesian Knowledge Tracing, knowledge graphs, mastery-based curricula, contextual bandits, and reinforcement learning to choose questions, feedback, or pacing. In other settings, online and continual-learning methods detect drift and update or select predictors using Bayesian updates, ensembles, few-shot adaptation, fast memory, or structural changes, with evaluations spanning tutoring, streaming forecasts, robotics, and dynamic graphs. A central challenge is balancing responsiveness to new conditions with retention, robustness, and computational cost, while demonstrating gains through longitudinal or shift-aware evaluation rather than short-term accuracy alone.
Latest papers
104 papers11 papers in the last four weeks, up 10% on the four weeks before. 0.1% of all new papers.