Improving feature interactions at Pinterest under industry constraints

TL;DR

This paper discusses the challenges and solutions for enhancing feature interactions in Pinterest's recommendation system within industry-specific constraints like latency and memory, providing practical insights for real-world applications.

Contribution

It offers a detailed account of how to adapt feature interaction techniques for industrial settings, including strategies, trade-offs, and experimental guidance.

Findings

Effective feature interaction strategies under industry constraints

Trade-offs between model performance and latency/memory limitations

Guidelines for selecting feature interaction architectures

Abstract

Adopting advances in recommendation systems is often challenging in industrial settings due to unique constraints. This paper aims to highlight these constraints through the lens of feature interactions. Feature interactions are critical for accurately predicting user behavior in recommendation systems and online advertising. Despite numerous novel techniques showing superior performance on benchmark datasets like Criteo, their direct application in industrial settings is hindered by constraints such as model latency, GPU memory limitations and model reproducibility. In this paper, we share our learnings from improving feature interactions in Pinterest's Homefeed ranking model under such constraints. We provide details about the specific challenges encountered, the strategies employed to address them, and the trade-offs made to balance performance with practical limitations. Additionally, we present a set of learning experiments that help guide the feature interaction architecture selection. We believe these insights will be useful for engineers who are interested in improving their model through better feature interaction learning.