DV365: Extremely Long User History Modeling at Instagram

TL;DR

This paper introduces DV365, a cost-effective method for modeling extremely long user histories up to 70,000 entries, significantly improving recommendation systems at Instagram with a novel embedding strategy.

Contribution

The paper presents a new multi-slicing and summarization user embedding strategy enabling long-term user interest modeling, integrated into Instagram's recommendation system.

Findings

Embedding length up to 70,000 entries

Significant impact on Instagram's recommendation models

Deployed in 15 models for over a year

Abstract

Long user history is highly valuable signal for recommendation systems, but effectively incorporating it often comes with high cost in terms of data center power consumption and GPU. In this work, we chose offline embedding over end-to-end sequence length optimization methods to enable extremely long user sequence modeling as a cost-effective solution, and propose a new user embedding learning strategy, multi-slicing and summarization, that generates highly generalizable user representation of user's long-term stable interest. History length we encoded in this embedding is up to 70,000 and on average 40,000. This embedding, named as DV365, is proven highly incremental on top of advanced attentive user sequence models deployed in Instagram. Produced by a single upstream foundational model, it is launched in 15 different models across Instagram and Threads with significant impact, and has been production battle-proven for >1 year since our first launch.