Behavior Sequence Transformer for E-commerce Recommendation in Alibaba

TL;DR

This paper introduces a Transformer-based model for e-commerce recommendations that captures user behavior sequences, outperforming traditional methods and improving online click-through rates in Alibaba's Taobao platform.

Contribution

The paper presents a novel Transformer model tailored for sequential user behavior data in e-commerce, demonstrating significant online performance improvements.

Findings

Superior online CTR compared to baseline models

Effective capture of sequential user behavior signals

Successful deployment in Alibaba's Taobao platform

Abstract

Deep learning based methods have been widely used in industrial recommendation systems (RSs). Previous works adopt an Embedding&MLP paradigm: raw features are embedded into low-dimensional vectors, which are then fed on to MLP for final recommendations. However, most of these works just concatenate different features, ignoring the sequential nature of users' behaviors. In this paper, we propose to use the powerful Transformer model to capture the sequential signals underlying users' behavior sequences for recommendation in Alibaba. Experimental results demonstrate the superiority of the proposed model, which is then deployed online at Taobao and obtain significant improvements in online Click-Through-Rate (CTR) comparing to two baselines.