Research map / Label-efficient and robust learning / Self-supervised and contrastive
NODE-GAM: Neural Generalized Additive Model for Interpretable Deep Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 NODE-GA²M instead of EBM-GA²M | GAMs on large datasets · Higgs | Error rate: 0.2767 -> 0.2566; 7.3% | Helped |
| use NODE-GA²M instead of EBM-GA²M | GAMs on large datasets · Microsoft | MSE: 0.5780 -> 0.5618; 2.8% | Helped |
| use NODE-GA²M instead of EBM-GA²M | GAMs on large datasets · Yahoo | MSE: 0.6032 -> 0.5807; 3.7% | Helped |
| use NODE-GA²M instead of EBM-GA²M | GAMs on large datasets · Year | MSE: 83.16 -> 79.57; 4.3% | Helped |
| use NODE-GA²M instead of EBM-GA²M | GAMs on large datasets · Click | Error rate: 0.3297 -> 0.3307; -0.2% | Hurt |
| use NODE-GA²M instead of EBM-GA²M | GAMs on medium-sized datasets · medium-sized datasets (average rank) | Rank: 3.5 -> 3.2 | Helped |
| use NODE-GA²M instead of EBM-GA²M | GAMs on medium-sized datasets · medium-sized datasets (normalized score) | NS: 0.812 -> 0.808 | No clear effect |
| add self-supervised masked-input pretraining (ablation) | NODE-GAM · six of seven limited-label datasets | AUC relative improvement over No-SS: 6% | Helped |
| add self-supervised masked-input pretraining (ablation) | NODE-GAM · Income with limited labeled data | AUC relative improvement over No-SS | No clear effect |
| add self-supervised masked-input pretraining (ablation) | NODE-GAM · Churn with limited labeled data | AUC relative improvement over EBM: up to 10% | Helped |
| use default rather than tuned hyperparameters (ablation) | NODE-GA²M · large datasets | error or MSE: 0.4%-3.7% | Hurt |
| use NODE-GA²M rather than NODE-GAM | interpretable models · large datasets including Higgs and Year | error rate / MSE: 0.2970 -> 0.2566; 85.09 -> 79.57 | Mixed |
Closest papers by meaning
- Additive Models Explained: A Computational Complexity Approach (NeurIPS 2025 · similarity 0.86)
- Gaussian Process Neural Additive Models (AAAI 2024 · similarity 0.86)
- Interpretable Generalized Additive Models for Datasets with Missing Values (NeurIPS 2024 · similarity 0.84)
- Interpretable Next-token Prediction via the Generalized Induction Head (NeurIPS 2025 · similarity 0.84)
- Mixture of Scope Experts at Test: Generalizing Deeper Graph Neural Networks with Shallow Variants (NeurIPS 2025 · similarity 0.83)
- To Grok or not to Grok: Disentangling Generalization and Memorization on Corrupted Algorithmic Datasets (ICLR 2024 · 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.