Allan, Zamani Co-Author Report on the Future of Information Retrieval in the Age of Generative AI
Computing Community Consortium workshop report offers a comprehensive examination of several crucial areas in information retrieval generative AI (IR-GenAI) research.
Content
The Computing Community Consortium (CCC) has released The Future of Information Retrieval Research in the Age of Generative AI Workshop Report, co-authored by Manning College of Information and Computer Sciences (CICS) Associate Dean of Research & Engagement and Director of the Center for Intelligent Information Retrieval James Allan and Associate Professor Hamed Zamani.
The report, written in collaboration with Eunsol Choi (New York University) and Daniel P. Lopresti (Lehigh University), builds on discussions from a workshop of the same name held during the summer of 2024 in Washington, D.C. It offers a comprehensive examination of several crucial areas in information retrieval generative AI (IR-GenAI) research and provides recommendations for academics, industry practitioners, institutions, evaluation campaigns, and funding agencies.
Each topic in the report is structured with a concise overview of its main observations, challenges, and opportunities, followed by short-term research recommendations of immediate progress and long-term recommendations that outline the broader trajectory of IR-GenAI research, including:
- Evaluation challenges and needs in IR-GenAI
- Learning from implicit and explicit human feedback for solving complex problems that may require reasoning
- Understanding and modeling users for the evolving generative AI-powered information access systems
- Challenges and potential solutions to address or mitigate socio-technical issues raised by the new technologies in IR-GenAI
- Methods for developing personalized IR-GenAI systems
- Efficiency considerations when scaling compute, data, and human efforts in developing IR-GenAI methods
- The role of information retrieval in enhancing AI agents
- Developing foundation models specifically for information access and discovery
Read the full workshop report on the Computing Community Consortium’s website; this article has been modified from a post originally published by the CCC.