UC San Diego

In today’s dynamic labour market, continuous re-skilling helps people adapt when changing jobs. Coaching can help people overcome plateaus when learning new skills that are needed for these new jobs. Yet, expert coaches are rare, and only a lucky few get to learn from them. How can we develop coaching systems and models that scale across a wide range of learning tasks? This thesis introduces Intelligent Coaching Systems (ICS). Inspired by human coaching, learning interactions, and AI techniques, an ICS evaluates observable outcomes of learners’ interactions with content and provides ambient feedback to cue learners’ attention to relevant resources. This thesis describes empirical work in three widely different domains, showcasing the domain-general potential of ICS. Lastly, this thesis introduces the ICS model that offers a blueprint for applying coaching in other learning settings, contributing to sustained and long-term learning.