Abstract:Modeling feature interactions is of crucial importance to predict click-through rate (CTR) in industrial recommender systems. However, manually crafting cross features usually requires extensive domain knowledge and labor-intensive feature engineering to obtain the desired cross features. To alleviate this problem, the factorization machine (FM) is proposed to model feature interactions from raw features automatically. In particular, it embeds each feature in a vector representation and discovers second-order interactions as the product of two feature representations. In order to learn nonlinear and complex patterns, recent works, such as NFM, PIN, and DeepFM, exploited deep learning techniques to capture higher-order feature interactions. These approaches lack guarantees about the effectiveness of high-order pattern as they model feature interactions in a rather implicit way. To address this limitation, xDeepFM is recently proposed to generate high-order interactions of features in an explicit fashion, where multiple interaction networks are stacked. Nevertheless, xDeepFM suffers from rather high complexity which easily leads to overfitting.In this paper, we develop a more expressive but lightweight solution based on FM, named High-order Attentive Factorization Machine (HoAFM), by accounting for the higher-order sparse feature interactions in an explicit manner. Beyond the linearity of FM, we devise a cross interaction layer, which updates a feature’s representation by aggregating the representations of other co-occurred features. In addition, we perform a bit-wise attention mechanism to determine the different importance of co-occurred features on the granularity of dimensions. By stacking multiple cross interaction layers, we can inject high-order feature interactions into feature representation learning, in order to establish expressive and informative cross features. Extensive experiments are performed on two benchmark datasets, Criteo and Avazu, to demonstrate the rationality and effectiveness of HoAFM. Empirical results suggest that HoAFM achieves significant improvement over other state-of-the-art methods, such as NFM and xDeepFM. https://github.com/zltao/HoAFM.

Input Enhanced Logarithmic Factorization Network for CTR Prediction

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

A Dual Adaptive Factorization Network for CTR Prediction.

A Dual Adaptive Interaction Click-Through Rate Prediction Based on Attention Logarithmic Interaction Network

Field-Aware Neural Factorization Machine for Click-Through Rate Prediction

FINAL:Factorized Interaction Layer for CTR Prediction

Memorize, Factorize, or be Naïve: Learning Optimal Feature Interaction Methods for CTR Prediction

Memorize, Factorize, or Be Naive: Learning Optimal Feature Interaction Methods for CTR Prediction

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

DeepFM: A Factorization-Machine based Neural Network for CTR Prediction

Towards Deeper, Lighter and Interpretable Cross Network for CTR Prediction

HoAFM: A High-order Attentive Factorization Machine for CTR Prediction

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

Feature Interaction based Neural Network for Click-Through Rate Prediction

Feature Interaction Fusion Self-Distillation Network For CTR Prediction

FLEN: Leveraging Field for Scalable CTR Prediction

Multi-Interactive Attention Network for Fine-grained Feature Learning in CTR Prediction

Context-aware cross feature attentive network for click-through rate predictions

A Click-Through Rate Prediction Method Based on Cross-Importance of Multi-Order Features

MCRF: Enhancing CTR Prediction Models Via Multi-channel Feature Refinement Framework

A Dual Input-aware Factorization Machine for CTR Prediction