Abstract:Click Through Rate (CTR) prediction plays an essential role in recommender systems and online advertising. It is crucial to effectively model feature interactions to improve the prediction performance of CTR models. However, existing methods face three significant challenges. First, while most methods can automatically capture high-order feature interactions, their performance tends to diminish as the order of feature interactions increases. Second, existing methods lack the ability to provide convincing interpretations of the prediction results, especially for high-order feature interactions, which limits the trustworthiness of their predictions. Third, many methods suffer from the presence of redundant parameters, particularly in the embedding layer. This paper proposes a novel method called Gated Deep Cross Network (GDCN) and a Field-level Dimension Optimization (FDO) approach to address these challenges. As the core structure of GDCN, Gated Cross Network (GCN) captures explicit high-order feature interactions and dynamically filters important interactions with an information gate in each order. Additionally, we use the FDO approach to learn condensed dimensions for each field based on their importance. Comprehensive experiments on five datasets demonstrate the effectiveness, superiority and interpretability of GDCN. Moreover, we verify the effectiveness of FDO in learning various dimensions and reducing model parameters. The code is available on \url{<a class="link-external link-https" href="https://github.com/anonctr/GDCN" rel="external noopener nofollow">this https URL</a>}.

What problem does this paper attempt to address?

The paper proposes a new approach to address three key challenges in Click-Through Rate (CTR) prediction: 1. **Effectiveness of High-Order Feature Interactions**: Many existing methods can automatically capture high-order feature interactions, but their performance often declines as the order of interactions increases. This is because not all high-order interactions are beneficial, leading to the introduction of unnecessary interactions, which reduces model performance and increases computational complexity. 2. **Interpretability**: Existing methods lack in explaining prediction results, especially in terms of high-order feature interactions, which limits the credibility of the prediction results. 3. **Parameter Redundancy**: Most existing models contain a large number of redundant parameters in the embedding layer, especially when all fields are assumed to have the same embedding dimension. To address these challenges, the paper proposes a new model called the "Gated Deep Cross Network" (GDCN) and a method called "Field-level Dimension Optimization" (FDO). - **GDCN**: Captures explicit high-order feature interactions through a Gated Cross Network (GCN) and dynamically filters important interactions through an information gate mechanism, reducing noise interference. Additionally, GCN is combined with a Deep Neural Network (DNN) to learn implicit feature interactions. - **FDO**: Allocates independent and condensed dimensions to each field based on its importance, effectively reducing the number of embedding parameters. Experimental results show that GDCN not only performs well on five datasets but also has good interpretability and generalization ability. FDO further improves the cost efficiency of the model, enabling GDCN to achieve faster training speeds while maintaining a smaller model size.

Towards Deeper, Lighter and Interpretable Cross Network for CTR Prediction

Click-Through Rate Prediction Algorithm Based on Modeling of Implicit High-Order Feature Importance

DCNv3: Towards Next Generation Deep Cross Network for CTR Prediction

Deep & Attentional Crossing Network for Click-through Rate Prediction

Graph Relation Embedding Network for Click-Through Rate Prediction

A Dual Adaptive Factorization Network for CTR Prediction.

An Efficient Deep Interaction Network for Click-Through Rate Prediction.

Dual Graph Enhanced Embedding Neural Network for CTR Prediction

Feature Generation by Convolutional Neural Network for Click-Through Rate Prediction

XCrossNet: Feature Structure-Oriented Learning for Click-Through Rate Prediction

CETN: Contrast-enhanced Through Network for CTR Prediction

Dynamic Parameterized Network for CTR Prediction

Neighbour Interaction based Click-Through Rate Prediction via Graph-masked Transformer

Feature Interaction Fusion Self-Distillation Network For CTR Prediction

Deep interaction network based CTR prediction model

Deep Position-wise Interaction Network for CTR Prediction

Enhancing Explicit and Implicit Feature Interactions via Information Sharing for Parallel Deep CTR Models

An Attention-based Deep Network for CTR Prediction

GFEN: Graph Feature Extract Network for Click-Through Rate Prediction.

A CTR Prediction Model with Double Matrix-Level Cross-Features

Coarse-to-fine: A dual-view attention network for click-through rate prediction