Search as a Mechanism for Learning

This NSF CAREER award project will examine how information retrieval (IR) systems are used for lifelong learning. While users commonly turn to the internet to learn about unfamiliar topics, the internet's algorithm-mediated IR systems, including search engines and conversational agents, have limitations when it comes to adequately supporting users with complex information needs. This is especially true when it comes to learning-oriented searches, with IR models unable to fully explain and predict how and what people search to learn. For example, motivations and metacognition are often neglected in representing the search intents during learning. Hence, relying on metrics like semantic relevance or past user behavior in IR systems may inadvertently result in suboptimal information experience of learners. Learn more.