Using Large Language Models (LLMs) for Building Recommender Systems: Opportunities and Challenges

LLMs have emerged as powerful tools for a wide range of NLP tasks. Recently, there has been a significant amount of interest in recommender system community in using LLMs to enhance various aspects of recommender systems. This talk will explain how LLMs are being incorporated in recommender systems, why should LLMs be used for building recommenders, and what are some of the associated challenges.

• Ways to utilize Generative LLMs for building recommender systems, including zero-shot, few-shot and fine-tuning methods.
• A review of the very recent research on this theme from academic and industrial labs.
• Understand the associated pros, cons, open questions and challenges of using LLMs for recommendations.

————————————————————————————————————————————————————

Sumit Kumar – Senior Machine Learning Engineer | Meta (Seattle ,USA)

I currently work as a senior MLE in the Recommendations domain at Meta. I have previously worked as an MLE in Recommendations domain at TikTok, as a Research Scientist at Alexa AI, Amazon, and as a Lead Software Engineer at Samsung R&D. I have over 10 years of work experience in software industry. I have taught professional courses at the University of Chicago and University of Washington. I love learning, teaching and building. I also run a blog and a newsletter on the Information Retrieval domain.