This paper presents a case study on deploying Large Language Models (LLMs) as an advanced "annotation" mechanism to achieve nuanced content understanding (e.g., discerning content "vibe") at scale within a large-scale industrial short-form video recommendation system. Traditional machine learning classifiers for content understanding face protracted development cycles and a lack of deep, nuanced comprehension. The "LLM-as-annotators" approach addresses these by significantly shortening development times and enabling the annotation of subtle attributes. This work details an end-to-end workflow encompassing: (1) iterative definition and robust evaluation of target attributes, refined by offline metrics and online A/B testing; (2) scalable offline bulk annotation of video corpora using LLMs with multimodal features, optimized inference, and knowledge distillation for broad application; and (3) integration of these rich annotations into the online recommendation serving system, for example, through personalized restrict retrieval. Experimental results demonstrate the efficacy of this approach, with LLMs outperforming human raters in offline annotation quality for nuanced attributes and yielding significant improvements of user participation and satisfied consumption in online A/B tests. The study provides insights into designing and scaling production-level LLM pipelines for rich content evaluation, highlighting the adaptability and benefits of LLM-generated nuanced understanding for enhancing content discovery, user satisfaction, and the overall effectiveness of modern recommendation systems.

Large Language Models (LLMs) empower recommendation systems through their advanced reasoning and planning capabilities. However, the dynamic nature of user interests and content poses a significant challenge: While initial fine-tuning aligns LLMs with domain knowledge and user preferences, it fails to capture such real-time changes, necessitating robust update mechanisms. This paper investigates strategies for updating LLM-powered recommenders, focusing on the trade-offs between ongoing fine-tuning and Retrieval-Augmented Generation (RAG). Using an LLM-powered user interest exploration system as a case study, we perform a comparative analysis of these methods across dimensions like cost, agility, and knowledge incorporation. We propose a hybrid update strategy that leverages the long-term knowledge adaptation of periodic fine-tuning with the agility of low-cost RAG. We demonstrate through live A/B experiments on a billion-user platform that this hybrid approach yields statistically significant improvements in user satisfaction, offering a practical and cost-effective framework for maintaining high-quality LLM-powered recommender systems.

Multi-task prediction models and value models are the de-facto standard ranking components in modern large-scale content recommendation systems. However, they are typically optimized to model users’ passive consumption behaviors, and rank content in a way to grow only consumption-centric values. In this talk, we discuss the key insight that it is possible to model sparse participatory content-generation actions as well and grow ecosystem value through a new ranking system. We made the following key technical contributions in this system: (1) introducing ranking for content generation based on a categorization of user participation actions of different sparsity, including proxy intent action or access point clicks. (2) improving sparse task prediction quality and stability by causal task relationship modeling, conditional loss modeling and ResNet based shared bottom network. (3) personalizing the value model to minimize conflicts between different values, through e.g. ranking inspiring content higher for users who actively generate content. (4) conducting systematic evaluation of proposed approach in a large short-form video UGC (User-Generated Content) platform.

Large volumes of spatio-temporal-thematic data being created using sites like Twitter and Jaiku, can potentially be combined to detect events, and understand various 'situations' as they are evolving at different spatio-temporal granularity across the world. Taking inspiration from traditional image pixels which represent aggregation of photon energies at a location, we consider aggregation of user interest levels at different geo-locations as social pixels. Combining such pixels spatio-temporally allows for creation of social images and video. Here, we describe how the use of relevant (media processing inspired) situation detection operators upon such 'images', and domain based rules can be used to decide relevant control actions. The ideas are showcased using a Swine flu monitoring application which uses Twitter data.

demonstration Share on MEDIALIFE: from images to a life chronicle Authors: Amarnath Gupta University of California San Diego, La Jolla, CA, USA University of California San Diego, La Jolla, CA, USAView Profile , Setareh Rafatirad University of California Irvine, Irvine, CA, USA University of California Irvine, Irvine, CA, USAView Profile , Mingyan Gao University of California Irvine, Irvine, CA, USA University of California Irvine, Irvine, CA, USAView Profile , Ramesh Jain University of California Irvine, Irvine, CA, USA University of California Irvine, Irvine, CA, USAView Profile Authors Info & Claims SIGMOD '09: Proceedings of the 2009 ACM SIGMOD International Conference on Management of dataJune 2009 Pages 1119–1122https://doi.org/10.1145/1559845.1559998Published:29 June 2009Publication History 4citation280DownloadsMetricsTotal Citations4Total Downloads280Last 12 Months2Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my Alerts New Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access