Research map / Learning theory and algorithms / Sample complexity
TAKDE: Temporal Adaptive Kernel Density Estimator for Real-Time Dynamic Density Estimation
Yinsong Wang, Yu Ding, Shahin Shahrampour
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 cross-validated smoothness parameter instead of normal bandwidth selector (ablation) | TAKDE · Wafer | mean test log-likelihood: 1.530 -> 1.848 | Improved |
| use cross-validated smoothness parameter instead of normal bandwidth selector (ablation) | TAKDE · TEM | mean test log-likelihood: 0.130 -> 0.362 | Improved |
| use cross-validated smoothness parameter instead of normal bandwidth selector (ablation) | TAKDE · ECG | mean test log-likelihood: 1.639 -> 1.648 | No clear effect |
| tune settings for TAKDE instead of KDEtrack's settings (ablation) | TAKDE · Earth | mean test log-likelihood: 1.331 -> 1.504 | Improved |
| use TAKDE with KDEtrack settings instead of KDEtrack | KDEtrack · TEM | mean test log-likelihood: 0.245 -> 0.246 | No clear effect |
| use AMISE-based weights instead of uniform or exponential weights (ablation) | TAKDE · synthetic Gaussian mixtures | test log-likelihood | Improved |
| increase cutoff value (memory) with heuristic weights (ablation) | uniform/exponential weighting · synthetic Gaussian mixtures | test log-likelihood | Hurt |
| increase number of batches from 100 to 500 (ablation) | TAKDE vs heuristic weights · synthetic Gaussian mixtures | test log-likelihood gap | Hurt |
| use TAKDE instead of B-spline Kalman Filter for speed | TAKDE · TEM | run-time (s): 7.08 -> 0.378 | Improved |
| use TAKDE instead of KDEtrack for speed | TAKDE · Star | run-time (s): 14.85 -> 0.851 | Improved |
| use TAKDE instead of B-spline Kalman Filter | TAKDE · Wafer | mean test log-likelihood: 1.204 -> 1.848 | Improved |
Closest papers by meaning
- SD-KDE: Score-Debiased Kernel Density Estimation (NeurIPS 2025 · similarity 0.87)
- Kernel Density Estimation for Dynamical Systems (JMLR 2018 · similarity 0.87)
- Uniform Convergence Rates for Kernel Density Estimation (ICML 2017 · similarity 0.87)
- Fast Exact Evaluation of Univariate Kernel Sums (TPAMI 2019 · similarity 0.86)
- Adaptive Greedy Algorithm for Moderately Large Dimensions in Kernel Conditional Density Estimation (JMLR 2022 · similarity 0.85)
- Overcoming Saturation in Density Ratio Estimation by Iterated Regularization (ICML 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.
