Boosting Factorization Machines via Saliency-Guided Mixup
Not on the map. Why?
Abstract (arXiv)
Factorization machines (FMs) are widely used in recommender systems due to their adaptability and ability to learn from sparse data. However, for the ubiquitous non-interactive features in sparse data, existing FMs can only estimate the parameters corresponding to these features via the inner product of their embeddings. Undeniably, they cannot learn the direct interactions of these features, which limits the model's expressive power. To this end, we first present MixFM, inspired by Mixup, to generate auxiliary training data to boost FMs. Unlike existing augmentation strategies that require labor costs and expertise to collect additional information such as position and fields, these extra data generated by MixFM only by the convex combination of the raw ones without any professional knowledge support. More importantly, if the parent samples to be mixed have non-interactive features, MixFM will establish their direct interactions. Second, considering that MixFM may generate redundant or even detrimental instances, we further put forward a novel Factorization Machine powered by Saliency-guided Mixup (denoted as SMFM). Guided by the customized saliency, SMFM can generate more informative neighbor data. Through theoretical analysis, we prove that the proposed methods minimize the upper bound of the generalization error, which hold a beneficial effect on enhancing FMs. Significantly, we give the first generalization bound of FM, implying the generalization requires more data and a smaller embedding size under the sufficient representation capability. Finally, extensive experiments on five datasets confirm that our approaches are superior to baselines. Besides, the results show that "poisoning" mixed data is likewise beneficial to the FM variants.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 44 of its 83 references that are in the map. First posted on arXiv 17 Jun 2022.
Builds on
44 of 83 references.
References · 44
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
- 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
- 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,371
- 2018DARTS: Differentiable Architecture Searchcited by 1,798
- 2020LightGCN: Simplifying and Powering Graph Convolution Network for Recommendationcited by 1,139
- 2015The Limitations of Deep Learning in Adversarial Settingscited by 790
- 2016Wide & Deep Learning for Recommender Systemscited by 747
- 2017DeepFM: A Factorization-Machine based Neural Network for CTR Predictioncited by 676
- 2016TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systemscited by 664
- 2017Deep & Cross Network for Ad Click Predictionscited by 393
- 2018Variational Autoencoders for Collaborative Filteringcited by 310
- 2018xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systemscited by 278
32 more
- 2020Attacks Which Do Not Kill Training Make Adversarial Learning Strongercited by 243
- 2020How Does Mixup Help With Robustness and Generalization?cited by 193
- 2017Neural Factorization Machines for Sparse Predictive Analyticscited by 183
- 2017Attentional Factorization Machines: Learning the Weight of Feature Interactions via Attention Networkscited by 126
- 2017Interpreting Deep Visual Representations via Network Dissectioncited by 103
- 2019FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Predictioncited by 83
- 2020On the Loss Landscape of Adversarial Training: Identifying Challenges and How to Overcome Themcited by 59
- 2020Uncertainty Inspired RGB-D Saliency Detectioncited by 56
- 2018Adversarial Personalized Ranking for Recommendationcited by 49
- 2018Improving the Generalization of Adversarial Training with Domain Adaptationcited by 48
- 2019Adaptive Factorization Network: Learning Adaptive-Order Feature Interactionscited by 48
- 2016Higher-Order Factorization Machinescited by 45
- 2021Learnable Embedding Sizes for Recommender Systemscited by 42
- 2021Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networkscited by 36
- 2021FM2: Field-matrixed Factorization Machines for Recommender Systemscited by 31
- 2022AutoField: Automating Feature Selection in Deep Recommender Systemscited by 30
- 2020MaxUp: A Simple Way to Improve Generalization of Neural Network Trainingcited by 27
- 2020Dropout: Explicit Forms and Capacity Controlcited by 27
- 2020On the Generalization Properties of Adversarial Trainingcited by 23
- 2019Predicting Different Types of Conversions with Multi-Task Learning in Online Advertisingcited by 21
- 2015Frappe: Understanding the Usage and Perception of Mobile App Recommendations In-The-Wildcited by 14
- 2021Automated Creative Optimization for E-Commerce Advertisingcited by 13
- 2021GraphFM: Graph Factorization Machines for Feature Interaction Modelingcited by 13
- 2022Single-shot Embedding Dimension Search in Recommender Systemcited by 13
- 2018Discrete Factorization Machines for Fast Feature-based Recommendationcited by 9
- 2020Towards Improved and Interpretable Deep Metric Learning via Attentive Groupingcited by 5
- 2015Multi-View Factorization Machinescited by 4
- 2018Android Malware Detection based on Factorization Machinecited by 2
- 2019RaFM: Rank-Aware Factorization Machinescited by 2
- 2017Learning from Multi-View Multi-Way Data via Structural Factorization Machinescited by 1
- 2018Contextual and Position-Aware Factorization Machines for Sentiment Classificationcited by 1
- 2022Additive Higher-Order Factorization Machinescited by 1
Built on by
Citing papers per year: 2022: 1; 2023: 0; 2024: 1; 2025: 0; 2026: 1.
3 papers.