and Search Systems
Recommendation and search systems play pivotal roles in how users access information, using them to discover news, entertainment, tutorials, housing, and employment, to name a few. As such, these systems influence social processes related to politics, culture, education, health, and economic well-being. The impact and risks across these systems are widely varied, from shaping the information consumed by users to uncertainty about what users want to challenges in simultaneously supporting a breadth of stakeholders. The potential for adverse impacts has resulted in increased attention from multiple stakeholders, including the academic community, policymakers, industry, and civil society.
This tutorial will cover four main topics: (1) content and experience quality, (2) bias and fairness, (3) diversity and filter bubbles, and (4) ecosystem effects. Each of these topics is complex and may be unfamiliar to many researchers and engineers designing these systems. However, there is a growing body of work in computer science and existing work in related disciplines that can inform the design of information access systems. For each, we will frame the set of concerns within each topic and then survey recent work for both measurement and modeling of these concerns. While each of these topics has been studied independently in the literature, we hope that by presenting them together we can give a more complete picture of how they interact and come together in recommendation and search system design and evaluation.
Preliminary Outline and References
A preliminary outline of our tutorial with references to the covered material can be found here.