User Intents and Journeys in Recommender Systems by Minmin Chen | VideoRecSys Workshop | RecSys 2023

"User Intents and Journeys in Recommender Systems" by Minmin Chen (Google Deepmind) at the Large-Scale Video Recommender Systems Workshop (VideoRecSys) 2023. In today's world, recommender systems have become an integral part of our digital lives, capturing users' interests based on their past interactions. However, these systems often produce user representations that are challenging to examine or interpret. What if we could understand user interests on a deeper, more personal level? In this presentation, Minmin Chen will introduce a novel framework designed to tackle the challenges of understanding user interests. The framework focuses on extracting and summarizing what we term "interest journeys" – those long-term user passions that deserve nuanced and personalized descriptions. To accomplish this, cutting-edge Large Language Models (LLMs) are employed, leveraging techniques like few-shot prompting, prompt-tuning and fine-tuning. The result is a more profound, interpretable and controllable understanding of users. By delving into the identification and comprehension of these extracted user journeys on a large-scale industrial platform, this talk talk is poised to pave a way toward recommendation platforms that possess an innate awareness of user journeys. For more information about this workshop, see our website: https://videorecsys.com #google #recsys #ml #ai