Search Engines that Learn from Their Users

Summary

This layman’s presentation from my PhD thesis defense addresses how search engines can improve their performance by learning from user interactions, a critical area given that over 500 million web searches occur daily and more than half the world’s population relies on search technology. The work presents a comprehensive framework spanning evaluation methods (including interleaving and multileaving techniques), methodology development through simulation frameworks, and experimentation with real users to understand and optimize search engine learning mechanisms. The research contributes both theoretical foundations and practical approaches for developing search systems that can adapt and improve based on user behavior patterns, addressing a fundamental challenge in information retrieval where search quality directly impacts billions of users worldwide.

On May 27, 2016, I publicly defended my dissertation Search Engines that Learn from Their Users at the University of Amsterdam.

My defense was recorded and was available online on the university website.