Research map / Label-efficient and robust learning / Recognition and long-tail
Interaction-Aware Factorization Machines for Recommender Systems
Fuxing Hong, Dongbo Huang, Ge Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add feature- and field-aware interaction importance | FM · Frappe | RMSE: 0.3321 -> 0.3080 | Improved |
| Add feature- and field-aware interaction importance | FM · MovieLens | RMSE: 0.4671 -> 0.4213 | Improved |
| Use the nonlinear interaction function G instead of IFM's sum of interaction matches | IFM · Frappe and MovieLens | RMSE: 0.3080 -> 0.3071; 0.4213 -> 0.4188 | Improved |
| Remove field-aspect interaction importance (ablation) | IFM · Frappe and MovieLens | RMSE | Hurt |
| Remove feature-aspect interaction importance (ablation) | IFM · Frappe and MovieLens | RMSE | Hurt |
| Sample 10% of feature interactions (ablation) | DeepIFM · Tencent video click-through prediction | AUC: 0.8436 -> 0.8420 | Hurt |
| Sample 10% of feature interactions (ablation) | DeepIFM · Tencent video click-through prediction | Time: 16hrs, 18mins -> 3hrs, 49mins | Improved |
| Factorize the field-aware factor importance matrix (ablation) | IFM · Frappe | RMSE | Improved |
| Factorize the field-aware factor importance matrix (ablation) | IFM · MovieLens | RMSE: around 0.1% | No clear effect |
| Apply L2 regularization to field-aspect factorization parameters (ablation) | IFM · MovieLens | RMSE: approximately 1.4% | Improved |
| Vary the dropout keep probability away from its best-tuned value (ablation) | IFM and FM · MovieLens and Frappe | RMSE | Mixed |
| Increase the number of field-aware hidden factors too far (ablation) | IFM · MovieLens and Frappe | RMSE | Hurt |
Closest papers by meaning
- Attention-over-Attention Field-Aware Factorization Machine (AAAI 2020 · similarity 0.90)
- Boosting Factorization Machines via Saliency-Guided Mixup (TPAMI 2024 · similarity 0.87)
- Detecting Beneficial Feature Interactions for Recommender Systems (AAAI 2021 · similarity 0.87)
- Factorization Machines with Regularization for Sparse Feature Interactions (JMLR 2021 · similarity 0.86)
- Learning Feature Interactions with Lorentzian Factorization Machine (AAAI 2020 · similarity 0.85)
- Improving LLMs for Recommendation with Out-Of-Vocabulary Tokens (ICML 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
