Higher-Order Factorization Machines

Abstract

Factorization machines (FMs) are a supervised learning approach that can use second-order feature combinations even when the data is very high-dimensional. Unfortunately, despite increasing interest in FMs, there exists to date no efficient training algorithm for higher-order FMs (HOFMs). In this paper, we present the first generic yet efficient algorithms for training arbitrary-order HOFMs. We also present new variants of HOFMs with shared parameters, which greatly reduce model size and prediction times while maintaining similar accuracy. We demonstrate the proposed approaches on four different link prediction tasks.

References in corpus (2)

• BPR: Bayesian Personalized Ranking from Implicit Feedback
• Tensor machines for learning target-specific polynomial features

Cited by in corpus (29)

• xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems
• Tensor Methods in Computer Vision and Deep Learning
• CatGCN: Graph Convolutional Networks with Categorical Node Features
• Disentangled Self-Attentive Neural Networks for Click-Through Rate Prediction
• Learning from Multi-View Multi-Way Data via Structural Factorization Machines
• Hyperbolic Deep Neural Networks: A Survey
• Adaptive Factorization Network: Learning Adaptive-Order Feature Interactions
• User Response Prediction in Online Advertising
• AutoCross: Automatic Feature Crossing for Tabular Data in Real-World Applications
• Deep Learning for Click-Through Rate Estimation
• Universal Approximation with Quadratic Deep Networks
• MiNet: Mixed Interest Network for Cross-Domain Click-Through Rate Prediction
• A Solution for Large Scale Nonlinear Regression with High Rank and Degree at Constant Memory Complexity via Latent Tensor Reconstruction
• Open Benchmarking for Click-Through Rate Prediction
• DNN2LR: Interpretation-inspired Feature Crossing for Real-world Tabular Data
• DeepLight: Deep Lightweight Feature Interactions for Accelerating CTR Predictions in Ad Serving
• Efficient Neural Interaction Function Search for Collaborative Filtering
• Iterative Boosting Deep Neural Networks for Predicting Click-Through Rate
• AEFE: Automatic Embedded Feature Engineering for Categorical Features
• Factorization Machines with Regularization for Sparse Feature Interactions
• Generalized Embedding Machines for Recommender Systems
• Enhancing Factorization Machines with Generalized Metric Learning
• Sequence-Aware Factorization Machines for Temporal Predictive Analytics
• Dynamic Parameterized Network for CTR Prediction
• XCrossNet: Feature Structure-Oriented Learning for Click-Through Rate Prediction
• Learning Feature Interactions with Lorentzian Factorization Machine
• DS-FACTO: Doubly Separable Factorization Machines
• Quaternion Factorization Machines: A Lightweight Solution to Intricate Feature Interaction Modelling
• DNN2LR: Automatic Feature Crossing for Credit Scoring