TransAct V2: Lifelong User Action Sequence Modeling on Pinterest Recommendation

TL;DR

TransAct V2 is a scalable, production-ready model that leverages long user action sequences and a novel loss function to improve CTR prediction and user action forecasting in large-scale recommendation systems.

Contribution

It introduces a new model architecture that effectively utilizes very long user sequences and integrates a Next Action Loss for better prediction accuracy in industrial settings.

Findings

Improved CTR prediction accuracy with long sequence modeling.

Enhanced user action forecasting through the Next Action Loss.

Efficient deployment solutions for large-scale sequential models.

Abstract

Modeling user action sequences has become a popular focus in industrial recommendation system research, particularly for Click-Through Rate (CTR) prediction tasks. However, industry-scale CTR models often rely on short user sequences, limiting their ability to capture long-term behavior. Additionally, these models typically lack an integrated action-prediction task within a point-wise ranking framework, reducing their predictive power. They also rarely address the infrastructure challenges involved in efficiently serving large-scale sequential models. In this paper, we introduce TransAct V2, a production model for Pinterest's Homefeed ranking system, featuring three key innovations: (1) leveraging very long user sequences to improve CTR predictions, (2) integrating a Next Action Loss function for enhanced user action forecasting, and (3) employing scalable, low-latency deployment solutions tailored to handle the computational demands of extended user action sequences.