Field-aware Factorization Machines for CTR Prediction

By Yu-Chin Juan, Yong Zhuang, Wei-Sheng Chin, Chih-Jen Lin. ACM Conference on Recommender Systems, 2016

797 citations recorded by the source corpus.

Subfields: Information Systems, Management Science and Operations Research, Marketing

View Paper: https://www.semanticscholar.org/paper/a802eccf56ea69c65b14831335a484a69e5cd849

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

1. Xiangnan He
2. Tat-seng Chua
3. Xiang Wang
4. Fuli Feng
5. Meng Wang
6. Liqiang Nie
7. Xing Xie
8. Hanwang Zhang

In Management Science and Operations Research

1. Xiangnan He
2. Xiang Wang
3. Tat-seng Chua
4. Meng Wang
5. Fuli Feng
6. Yifei Ma
7. Balakrishnan Narayanaswamy
8. Tongwen Huang

In Marketing

1. Eric P. Xing
2. Tat-seng Chua
3. Fei Wu
4. Jun Xiao
5. Jian Tang
6. Yong Yu
7. Xiuqiang He
8. Yu-Xiang Wang

Citations in the corpus

• Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networks (2017, background): “…as the neural FM [He and Chua, 2017] that deepens FM under the neural framework to learn high-order feature interactions, and the field-aware FM [ Juan et al. , 2016 ] that associates multiple embedding vectors for a feature to differentiate its interaction with other features of different…”
• Neural Factorization Machines for Sparse Predictive Analytics (2017, background): “It has been successfully applied to many predictive tasks, ranging from online advertising [21], microblog retrieval [26], to open relation extraction [25].”
• Neural Graph Collaborative Filtering (2019, background): “Personalized recommendation is ubiquitous, having been applied to many online services such as E-commerce [43] and advertising [20].”
• Breaking the Nonsmooth Barrier: A Scalable Parallel Method for Composite Optimization (2017, neutral): “15 ) achieves a 11x speedup, while in Criteo ( ∆ = 0 .”
• Towards Automated Neural Interaction Discovery for Click-Through Rate Prediction (2020, background): “Since its inception, the mainstream models have roughly experienced three-stage evolution starting from linear regression and tree-based models [11, 20, 32], to interaction-based models [14, 26], and then the deep neural networks (DNNs) [7, 9, 10, 16, 19, 21, 23, 30, 35, 36].”
• Field-weighted Factorization Machines for Click-Through Rate Prediction in Display Advertising (2018, methodology): “Recently, FFMs have been among the best performing models for CTR prediction and won two competitions hosted by Criteo and Avazu [13].”
• Deep & Cross Network for Ad Click Predictions (2017, background): “Field-aware factorization machines (FFMs) [7, 8] further allow each feature to learn several vectors where each vector is associated with a €eld.”
• Imitation-Regularized Offline Learning (2019, background): “Formally, we estimate Q(a, x), the expected reward from taking an action a in context x; decisions are then implied by the greedy policy that selects actions with the highest expected reward in each decision context [15, 23, 11, 21].”
• AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks (2018, background): “For example, Field-aware Factorization Machines (FFM) [16] modeled fine-grained interactions between features from different fields.”
• FiBiNET: combining feature importance and bilinear feature interaction for click-through rate prediction (2019, methodology): “The shallow baseline models include LR(logistic regression)[17], FM[19, 20], FFM[9, 10], AFM[23], and the deep baseline models include FNN[25], DCN[22], DeepFM[4], XDeepFM[15].”
• AutoGroup: Automatic Feature Grouping for Modelling Explicit High-Order Feature Interactions in CTR Prediction (2020, methodology): “For sparse data, Factorization Machine (FM) and its variants [24, 29] project each feature into a low-dimensional vector and models feature interactions by inner product while FFM [14] enables each feature to have multiple latent vectors to interact with features from different fields.”
• Fault Tolerance in Iterative-Convergent Machine Learning (2018, methodology): “2xlarge instances to train MLR on the full 26GB Criteo (Juan et al., 2016) dataset, and LDA on a 12GB subset of the ClueWeb12 dataset (Gabrilovich et al.”
• Practical Privacy Preserving POI Recommendation (2020, background): “, DeepFM [24] and Field-aware FM [27], are also extensively used since they can capture the high-order interactions between features.”
• DHEN: A Deep and Hierarchical Ensemble Network for Large-Scale Click-Through Rate Prediction (2022, background): “FFMs [16, 17] extends FM, by assigning each feature with multiple vectors to interact with features from other fields.”