PinnerSage: Multi-Modal User Embedding Framework for Recommendations at Pinterest

Abstract

Latent user representations are widely adopted in the tech industry for powering personalized recommender systems. Most prior work infers a single high dimensional embedding to represent a user, which is a good starting point but falls short in delivering a full understanding of the user's interests. In this work, we introduce PinnerSage, an end-to-end recommender system that represents each user via multi-modal embeddings and leverages this rich representation of users to provides high quality personalized recommendations. PinnerSage achieves this by clustering users' actions into conceptually coherent clusters with the help of a hierarchical clustering method (Ward) and summarizes the clusters via representative pins (Medoids) for efficiency and interpretability. PinnerSage is deployed in production at Pinterest and we outline the several design decisions that makes it run seamlessly at a very large scale. We conduct several offline and online A/B experiments to show that our method significantly outperforms single embedding methods.

10 pages, 7 figures

References in corpus (7)

• Distributed Representations of Words and Phrases and their Compositionality
• Graph Convolutional Neural Networks for Web-Scale Recommender Systems
• Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs
• Is a Single Embedding Enough? Learning Node Representations that Capture Multiple Social Contexts
• Is a Single Vector Enough? Exploring Node Polysemy for Network Embedding
• Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time
• Learning Item-Interaction Embeddings for User Recommendations

Cited by in corpus (15)

• ETA Prediction with Graph Neural Networks in Google Maps
• Combinatorial Optimization with Physics-Inspired Graph Neural Networks
• Trustworthy Graph Neural Networks: Aspects, Methods and Trends
• TextGNN: Improving Text Encoder via Graph Neural Network in Sponsored Search
• TwHIN: Embedding the Twitter Heterogeneous Information Network for Personalized Recommendation
• Ginex: SSD-enabled Billion-scale Graph Neural Network Training on a Single Machine via Provably Optimal In-memory Caching
• Accelerating Sampling and Aggregation Operations in GNN Frameworks with GPU Initiated Direct Storage Accesses
• ECLARE: Extreme Classification with Label Graph Correlations
• RoarGraph: A Projected Bipartite Graph for Efficient Cross-Modal Approximate Nearest Neighbor Search
• Representation Online Matters: Practical End-to-End Diversification in Search and Recommender Systems
• OmniSage: Large Scale, Multi-Entity Heterogeneous Graph Representation Learning
• kNN-Embed: Locally Smoothed Embedding Mixtures For Multi-interest Candidate Retrieval
• Understanding the Design Principles of Link Prediction in Directed Settings
• Securing Visually-Aware Recommender Systems: An Adversarial Image Reconstruction and Detection Framework
• Suggest, Complement, Inspire: Story of Two Tower Recommendations at Allegro.com