Siamese Cookie Embedding Networks for Cross-Device User Matching

TL;DR

This paper introduces SCEmNet, a supervised siamese convolutional network that improves cross-device user matching by leveraging multi-modal user event sequences, advancing beyond previous unsupervised methods.

Contribution

The paper presents a novel supervised siamese convolutional architecture, SCEmNet, for cross-device user matching that outperforms existing state-of-the-art approaches.

Findings

SCEmNet significantly improves matching accuracy.

Supervised learning outperforms unsupervised methods.

Multi-modal sequence integration enhances performance.

Abstract

Over the last decade, the number of devices per person has increased substantially. This poses a challenge for cookie-based personalization applications, such as online search and advertising, as it narrows the personalization signal to a single device environment. A key task is to find which cookies belong to the same person to recover a complete cross-device user journey. Recent work on the topic has shown the benefits of using unsupervised embeddings learned on user event sequences. In this paper, we extend this approach to a supervised setting and introduce the Siamese Cookie Embedding Network (SCEmNet), a siamese convolutional architecture that leverages the multi-modal aspect of sequences, and show significant improvement over the state-of-the-art.