Call for Participation: SIGIR 2024 Large Language Model Day

July 16, 2024 - Washington DC, USA

Large Language Models (LLMs) have become increasingly central to search and recommendation scenarios due to their remarkable capabilities in natural language understanding and generation. Examples of LLMs include commercial products such as GPT-4 and Gemini, as well as open-source models such as LLaMA and Mistral. These models are demonstrating great potential to revolutionize various aspects of search and recommendation systems – including core algorithms, system architecture, content understanding, user understanding, interactive modalities, domain-specific applications, and how such systems impact society.

As the focus on LLMs have increased in all fields, the rate of innovation across the world has also accelerated. SIGIR 2024 will feature a Large Language Model Day (LLM Day) as a themedaddition to the SIGIR main conference We embrace LLMs into IR research and complement the offerings of the main conference schedule. The LLM Day will facilitate a timely discussion on recent research and applications of LLMs covering both the core SIGIR community as well as other communities. In addition, attendees of LLM Day will receive a schedule highlighting relevant talks on LLMs throughout the conference to help navigate the full conference.

LLM Day focuses on the following types of work:

  • Recent progress of foundational research in LLMs in academia and industry;
  • Emerging applications of LLMs;
  • Innovating and building on open-source LLMs;
  • Intersection of core IR research and LLMs, including IR for LLMs and LLMs for IR;
  • The state-of-practice of LLMs for search and recommendation;
  • Evaluation of foundation models.

We expect the likely topics to include but not be limited to: RAG and Dense Retrieval, Inference Efficiency, Alignment and Fine-tuning from Open-Ended Feedback, Generating Synthetic Data, Generative Retrieval and Recommendations, Long-Context Understanding, Multimodal Interaction, and AI Ethics & Safety.

Registration for attendees is open for all interested in the topic. Speakers and panelists for LLM Day will be decided by invitation. These talks or panels will represent impactful and emerging research and applications from a variety of top-tier venues as well as industry. In order to focus on the most recent trends and research, we will announce the detailed schedule for LLM day later.


Large Language Model Day Chairs

  • Yue Wang, University of North Carolina at Chapel Hill, USA
  • Paul Bennett, Spotify, USA
  • Julia Kiseleva, Microsoft Research, USA

Contact

Questions or suggestions related to LLM Day can be communicated to the SIGIR 2024 LLM Day Chairs at https://forms.gle/hUmJGt6xgu9N5xPB6 (Google Form) or sigir24-llmday@acm.org