InterFormer: Effective Heterogeneous Interaction Learning for Click-Through Rate Prediction

TL;DR

InterFormer introduces a bidirectional, interleaving approach to learn heterogeneous user information for improved click-through rate prediction, overcoming limitations of previous unidirectional and aggressive aggregation methods.

Contribution

The paper presents InterFormer, a novel module enabling bidirectional interaction and effective information retention for CTR prediction, advancing beyond prior unidirectional and early summarization techniques.

Findings

Achieves state-of-the-art results on three public datasets.

Demonstrates superior performance on a large-scale industrial dataset.

Effectively models heterogeneous information interactions in CTR prediction.

Abstract

Click-through rate (CTR) prediction, which predicts the probability of a user clicking an ad, is a fundamental task in recommender systems. The emergence of heterogeneous information, such as user profile and behavior sequences, depicts user interests from different aspects. A mutually beneficial integration of heterogeneous information is the cornerstone towards the success of CTR prediction. However, most of the existing methods suffer from two fundamental limitations, including (1) insufficient inter-mode interaction due to the unidirectional information flow between modes, and (2) aggressive information aggregation caused by early summarization, resulting in excessive information loss. To address the above limitations, we propose a novel module named InterFormer to learn heterogeneous information interaction in an interleaving style. To achieve better interaction learning, InterFormer enables bidirectional information flow for mutually beneficial learning across different modes. To avoid aggressive information aggregation, we retain complete information in each data mode and use a separate bridging arch for effective information selection and summarization. Our proposed InterFormer achieves state-of-the-art performance on three public datasets and a large-scale industrial dataset.