SIGIR Day 1 - Keynotes and Industry Papers

Day 1 started with the opening remarks from general/program chairs. Some key insights are as follows:

• RecSys has the highest number of submissions for both full papers (458) and short papers (164), indicating a significant interest in this topic.
• Multi Modal IR and Search and Ranking are other prominent topics with substantial submissions and acceptance rates.

• Major tech companies like Microsoft, Google, and Amazon continue to play a pivotal role in research, with significant paper acceptances.
• The list highlights a broad international presence, with leading institutions from China, Europe, and the USA contributing extensively to the conference.

• GenAI played a crucial part in the review process

• Keynote 1: Ellen Vorhees, recipient of Salton award this year gave an interesting talk on her journey focused on IR evaluation (TREC)

• Keynote 2: Prof. Yimin Yang from CMU presented Representation Learning and Information Retrieval - Her lab is doing very interesting on easy to hard problem learning in LLMs

How to obtain more confident results from LLMs?

• SIRIP (IR in Practice) - Keynote 1 - Paul Bennett from Spotify gave an enlightening talk on "Toward AI-Powered Next Generation Personalized Experiences"
• LLM embeddings (content understanding) + GNNs (co-listenings) => user-item model
• The integration of GNNs with LLMs for content and consumption information across formats (like audiobooks and podcasts) has demonstrated significant performance improvements.

• SIRIP: Ferhan Ture from Comcast presented: "Ask Me Anything": How Comcast Uses LLMs to Assist Agents in Real Time

• SIRIP: Folks from Microsoft presented- A Field Guide to Automatic Evaluation of LLM-Generated Summaries (very insightful)

• SIRIP: Folks from LinkedIn presented: Retrieval-Augmented Generation with Knowledge Graphs for Customer Service Question Answering

• SIRIP: Amazon music group presented their work on "A Comprehensive Approach to Amazon Music Search Spell Correction"

There were many other interesting work worth highlighting and discussing but I do not intend to list all of them here. Please look at this: https://sigir-2024.github.io/program_SIRIP.html