Research map / Label-efficient and robust learning / Recognition and long-tail
Learning Feature Interactions with Lorentzian Factorization Machine
Canran Xu, Ming Wu
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 |
|---|---|---|---|
| Use Lorentz triangle-inequality feature interaction pooling instead of conventional deep feature-interaction layers | LorentzFM · Steam | MRR: 0.1284 -> 0.1362 | Improved |
| Use Lorentz triangle-inequality feature interaction pooling instead of conventional deep feature-interaction layers | LorentzFM · KKBox | MRR: 0.0644 -> 0.0733 | Improved |
| Use Lorentz triangle-inequality feature interaction pooling instead of conventional deep feature-interaction layers | LorentzFM · Avazu | AUC: 0.7677 -> 0.7775 | Improved |
| Use Lorentz triangle-inequality feature interaction pooling instead of conventional deep feature-interaction layers | LorentzFM · MovieLens | MRR: 0.0791 -> 0.0780 | Hurt |
| Increase embedding size | LorentzFM · Steam | performance | Improved |
| Remove the deep learning layers above feature interactions by using LorentzFM's lightweight architecture | LorentzFM · MovieLens | training parameters: 82.9% | Improved |
| Remove the deep learning layers above feature interactions by using LorentzFM's lightweight architecture | LorentzFM · MovieLens | training time per epoch: 69.7% | Improved |
| Use LorentzFM's lightweight architecture | LorentzFM · Steam | training time per epoch: -8.1% | Hurt |
Closest papers by meaning
- Interaction-Aware Factorization Machines for Recommender Systems (AAAI 2019 · similarity 0.85)
- Holographic Factorization Machines for Recommendation (AAAI 2019 · similarity 0.84)
- Attention-over-Attention Field-Aware Factorization Machine (AAAI 2020 · similarity 0.83)
- REACTION: Parameter-Efficient Learning for Recommendation (AAAI 2026 · similarity 0.83)
- From Feature Interaction to Feature Generation: A Generative Paradigm of CTR Prediction Models (ICML 2025 · similarity 0.82)
- Hgformer: Hyperbolic Graph Transformer for Collaborative Filtering (ICML 2025 · similarity 0.81)
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.
