The report linked below was published by the Computing Research Association (CRA)’s Computing Community Consortium (CCC).
title
The future of information retrieval research in the era of generative AI
author
james allan
University of Massachusetts Amherst
Choi Eun Seol
University of Texas at Austin
new york university
Daniel P. Lopresti
lehigh university
CCC
Hamed Zamani
University of Massachusetts Amherst
sauce
CCC via arXiv
abstract
executive summary
In the rapidly evolving field of information retrieval (IR), the integration of generative AI technologies such as large-scale language models (LLMs) is changing the way users search for and interact with information. Recognizing this paradigm shift at the intersection of IR and generative AI (IR-GenAI), a vision workshop supported by the Computing Community Consortium (CCC) will be held in July 2024 to discuss the future of IR in the era of generative AI. It was done. The workshop brought together 44 experts in information retrieval, natural language processing, human-computer interaction, and artificial intelligence from academia, industry, and government to discuss how generative AI can enhance IR and We explored the opposite possibility and identified key challenges and opportunities in this issue. A rapidly evolving field.
workshop activities
The workshop began with an overview of previous discussions from preliminary brainstorming sessions and other related workshops, and included presentations from selected participants to spark deeper discussion. To explore specific themes, eight breakout sessions were formed to discuss different aspects of the topic. The workshop agenda can be found on the CCC website 1.
Mechanisms for effective collaboration
To collaborate effectively on discussion topics, we followed IDEO’s brainstorming methodology (Seven Simple Rules for Brainstorming, n.d.), which is based on the following seven mechanisms for effective collaboration: (1 ) withhold judgment, (2) encourage out-of-the-box ideas, (3) ) build on others’ ideas, (4) stay focused on the topic, and (5) one conversation at a time. (6) (7) Discuss as many ideas as possible (focus on quantity of ideas), do so visually if necessary.
Research direction
This report outlines eight research directions for IR-GenAI systems that have high intellectual merit and far-reaching impact. (1) Evaluation issues and needs in IR-GenAI. (2) learn from implicit and explicit human feedback to solve complex problems that require reasoning; (3) Understand and model users of evolving generative AI-powered information access systems. (4) challenges and potential solutions to address or alleviate the socio-technical problems posed by new technologies in IR-GenAI; (5) How to develop a personalized IR-GenAI system. (6) Efficiency considerations in scaling compute, data, and human effort in developing IR-GenAI methods. (7) The role of information retrieval in enhancing AI agents. (8) Developing foundational models specific to information access and discovery.
The report includes an overview of the debate as a potentially important research topic and includes a list of recommendations for academics, industry participants, institutions, evaluation campaigns, and funding agencies.
Direct to full-text report
39 pages. PDF.
Files: Data Files, Funding, News, Patrons and Users
