Research map / Learning theory and algorithms / Prediction and decision losses
Sparse Interaction Additive Networks via Feature Interaction Detection and Sparse Selection
James Enouen, Yan Liu
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 |
|---|---|---|---|
| Increase SIAN interaction order from 1 to 5 (ablation) | SIAN · Appliances Energy | MSE: 0.718±0.007 -> 0.801±0.031 | Hurt |
| Increase SIAN interaction order from 1 to 3 (ablation) | SIAN · Bike Sharing | MSE: 0.387±0.035 -> 0.125±0.013 | Improved |
| Increase SIAN interaction order from 1 to 5 (ablation) | SIAN · California Housing | MSE: 0.378±0.007 -> 0.272±0.003 | Improved |
| Increase SIAN interaction order from 1 to 5 (ablation) | SIAN · Wine Quality | MSE: 0.551±0.004 -> 0.484±0.006 | Improved |
| Increase SIAN interaction order from 1 to 5 (ablation) | SIAN · Song Year | MSE: 0.860±0.001 -> 0.821±0.001 | Improved |
| Increase SIAN interaction order from 1 to 5 (ablation) | SIAN · Higgs Boson | AUROC: 0.771±0.001 -> 0.802±0.001 | Improved |
| Extend SIAN interaction order from 2 to 3 (ablation) | SIAN · MIMIC-III | AUROC: 0.855 -> 0.856 | Improved |
| Extend NODE-GA²M to trivariate interactions | NODE-GA²M · California Housing | MSE: 0.222 to 0.175 | Improved |
| Extend NODE-GA²M to trivariate interactions | NODE-GA²M · Wine Quality | MSE: 0.521 to 0.496 | Improved |
| Use block-sparse parallel computation instead of default sequential shape-function computation (ablation) | SIAN-2 · Wine Quality | Training time (minutes): 17.57 -> 0.69 | Improved |
| Save parameters in compressed mode (ablation) | SIAN-2 · Wine Quality | Size on disk (KB): 7,113 -> 537 | Improved |
Closest papers by meaning
- Towards Hybrid-grained Feature Interaction Selection for Deep Sparse Network (NeurIPS 2023 · similarity 0.90)
- Progressive Feature Interaction Search for Deep Sparse Network (NeurIPS 2021 · similarity 0.87)
- Factorization Machines with Regularization for Sparse Feature Interactions (JMLR 2021 · similarity 0.86)
- Provable Explanations for Any-Order Neural Additive Models (NeurIPS 2026 · similarity 0.85)
- Monitoring Primitive Interactions During the Training of DNNs (AAAI 2025 · similarity 0.84)
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.84)
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.
