CAREER: Search as a Mechanism for Learning

About This Grant

As our society and technology continue to evolve rapidly, there is a growing interests by adults in taking an active role to continue to learn on their own as hey age. One of the most common lifelong learning scenarios is gaining knowledege from an algorithm-mediated information environment, such as using the internet to explore unfamiliar topics or interact with conversational agents to adopt new health behaviors. Contemporary algorithm-mediated information retrieval (IR) systems, including search engines and conversational agents, have limitations when it comes to adequately supporting users with complex information needs, particularly those related to learning-oriented search tasks. Finding information does not necessarily lead to effective learning or deep comprehension. Moreover, IR algorithms often neglect users needs or intentions for the search - one search does not fit all. Measures of intent and algorithms need to be tuned to individual users or risk misaligning with the learning goals of users. Therefore, developing IR systems for learning requires understanding how individuals monitor, assess, and regulate their learning progress and what factors shape their judgments to persist in or disengage from learning. To facilitate the translation of research into instructional and outreach practices, the project will collaborate with the Osher Lifelong Learning Institution and the National Multiple Sclerosis Society to co-design educational games and webinars that foster information literacy and its practical applications in everyday search and learning contexts. The project seeks to advance the understanding of self-regulated learning as mediated by information search behavior and IR systems, while also pioneering the development of an intelligent system that scaffolds self-regulated learning through adaptive information search. The project is grounded in an interdisciplinary theoretical framework that integrates cognitive science theories, specifically information foraging, metacognition, self-regulated learning, and comprehension, with the capacities of large language models. First, the project examines the cognitive mechanisms underlying learning through information search, aiming to develop generalizable user models that characterize the cues individuals use to assess learning and how users regulate their resources between exploitation and exploration to optimize learning outcomes. Second, the project expands user models by considering contextual factors from diverse IR systems, including search engines and conversational agents, across the lifespan. Third, the project employs user models, symbolic knowledge representation, and large language models to develop intelligent systems that support self-regulated learning through information search. Overall, the overarching goal of the project is to empower adults across the lifespan to adapt to evolving information environments for lifelong learning. The research will guide the development of personalized technologies that promote lifelong learning in various educational settings, including online adult vocational education and individualized tutoring systems for all learners. The research outcomes and approach can be translated to other domains, enabling proactive prediction of user behavior and the delivery of tailored information experiences to enhance cognitive performance or health behaviors for adults across the lifespan. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.