Deep Interest Network for Click-Through Rate Prediction

Abstract

Click-through rate prediction is an essential task in industrial applications, such as online advertising. Recently deep learning based models have been proposed, which follow a similar Embedding\&MLP paradigm. In these methods large scale sparse input features are first mapped into low dimensional embedding vectors, and then transformed into fixed-length vectors in a group-wise manner, finally concatenated together to fed into a multilayer perceptron (MLP) to learn the nonlinear relations among features. In this way, user features are compressed into a fixed-length representation vector, in regardless of what candidate ads are. The use of fixed-length vector will be a bottleneck, which brings difficulty for Embedding\&MLP methods to capture user's diverse interests effectively from rich historical behaviors. In this paper, we propose a novel model: Deep Interest Network (DIN) which tackles this challenge by designing a local activation unit to adaptively learn the representation of user interests from historical behaviors with respect to a certain ad. This representation vector varies over different ads, improving the expressive ability of model greatly. Besides, we develop two techniques: mini-batch aware regularization and data adaptive activation function which can help training industrial deep networks with hundreds of millions of parameters. Experiments on two public datasets as well as an Alibaba real production dataset with over 2 billion samples demonstrate the effectiveness of proposed approaches, which achieve superior performance compared with state-of-the-art methods. DIN now has been successfully deployed in the online display advertising system in Alibaba, serving the main traffic.

Accepted by KDD 2018

References in corpus (9)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Densely Connected Convolutional Networks
• Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
• DeepFM: A Factorization-Machine based Neural Network for CTR Prediction
• Wide & Deep Learning for Recommender Systems
• DKN: Deep Knowledge-Aware Network for News Recommendation
• ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation
• Learning Piece-wise Linear Models from Large Scale Data for Ad Click Prediction
• Product-based Neural Networks for User Response Prediction

Cited by in corpus (22)

• xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems
• Searching for Activation Functions
• Learning Tree-based Deep Model for Recommender Systems
• DKN: Deep Knowledge-Aware Network for News Recommendation
• ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation
• Compositional Embeddings Using Complementary Partitions for Memory-Efficient Recommendation Systems
• Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate
• Learning from Multi-View Multi-Way Data via Structural Factorization Machines
• Image Matters: Visually modeling user behaviors using Advanced Model Server
• MARS: Memory Attention-Aware Recommender System
• DADNN: Multi-Scene CTR Prediction via Domain-Aware Deep Neural Network
• Layer-wise Relevance Propagation for Explainable Recommendations
• DeepLight: Deep Lightweight Feature Interactions for Accelerating CTR Predictions in Ad Serving
• Feature Interaction based Neural Network for Click-Through Rate Prediction
• Muddling Label Regularization: Deep Learning for Tabular Datasets
• Scaling Up Collaborative Filtering Data Sets through Randomized Fractal Expansions
• Future-Aware Diverse Trends Framework for Recommendation
• A Hierarchical User Intention-Habit Extract Network for Credit Loan Overdue Risk Detection
• Dynamic Parameterized Network for CTR Prediction
• Frequency-aware SGD for Efficient Embedding Learning with Provable Benefits
• Low-Precision Hardware Architectures Meet Recommendation Model Inference at Scale
• Itinerary-aware Personalized Deep Matching at Fliggy