Research map / Time series / Causal and generative time series
Curve Your Enthusiasm: Concurvity Regularization in Differentiable Generalized Additive Models
Julien Niklas Siems, Konstantin Ditschuneit, Winfried Ripken, Alma Lindborg, Maximilian Schambach, Johannes Otterbach, Martin Genzel
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 concurvity regularizer | NeuralProphet with many Fourier terms · synthetic daily+weekly seasonality | correlation between weekly and daily components: -0.64 -> 5.2 · 10 | Improved |
| add concurvity regularizer | NeuralProphet with many Fourier terms · synthetic daily+weekly seasonality | RMSE: 4.68 · 10-3 -> 4.37 · 10-3 | No clear effect |
| use many Fourier terms instead of NeuralProphet defaults | NeuralProphet · synthetic daily+weekly seasonality | RMSE: 3.85 · 10-1 -> 4.68 · 10-3 | Improved |
| add concurvity regularizer | NAM · non-linear toy example | feature correlation of f1,f2: 0.92 -> 0.03 | Improved |
| add concurvity regularizer | NAM · California Housing | concurvity R_perp: 0.2 -> 0.05 | Improved |
| add concurvity regularizer | NAM · California Housing | validation RMSE: 0.59 -> 0.66 | Hurt |
| add concurvity regularizer (lambda=0.1) | NAM · California Housing, 60 seeds | mean test RMSE: 0.58 -> 0.64 | Hurt |
| add concurvity regularizer | NAM · toy example of [28] | concurvity R_perp: 0.22 -> 0.03 | Improved |
| add concurvity regularizer | NAM · toy example of [28] | R2 (%): 80.77 -> 79.28 | Hurt |
| use L1 regularization instead of concurvity regularization (ablation) | NAM · toy example of [28] | concurvity R_perp: 0.03 -> 0.21 | Hurt |
| add L1 regularization (ablation) | NAM · toy example of [28] | R2 (%): 80.77 -> 79.12 | Hurt |
| add concurvity or L1 regularization | NAM · toy example of [28] | feature selection (true features X1, X5, X6) | Improved |
Closest papers by meaning
- Fast Automatic Smoothing for Generalized Additive Models (JMLR 2019 · similarity 0.86)
- Additive Models Explained: A Computational Complexity Approach (NeurIPS 2025 · similarity 0.85)
- Interpretable Generalized Additive Models for Datasets with Missing Values (NeurIPS 2024 · similarity 0.84)
- Position: Understanding LLMs Requires More Than Statistical Generalization (ICML 2024 · similarity 0.83)
- Beyond Log-Concavity and Score Regularity: Improved Convergence Bounds for Score-Based Generative Models in W2-distance (ICML 2025 · similarity 0.83)
- Spurious Correlations in High Dimensional Regression: The Roles of Regularization, Simplicity Bias and Over-Parameterization (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.
