Research map / Learning theory and algorithms / Sample complexity
SURF: A Simple, Universal, Robust, Fast Distribution Learning Algorithm
Yi Hao, Ayush Jain, Alon Orlitsky, Vaishakh Ravindrakumar
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 SURF (no t needed) instead of ADLS with a misspecified number of pieces t | piecewise-linear estimator · Beta, Gamma, Gaussian mixture synthetic | l1 error | Improved |
| Use ADLS with t=5, roughly the number of pieces SURF outputs | ADLS · Beta, Gamma, Gaussian mixture synthetic | l1 error | No clear effect |
| Increase number of samples n | SURF d=1,2,3 · Beta mixtures f1,f2,f3 | l1 error | Improved |
| Increase polynomial degree d from 1 to 3 (ablation) | SURF · Beta mixtures f1,f2,f3 | l1 error across distributions | Mixed |
| Increase n from 1024 to 65536 | SURF d=2 · 0.3Beta(3,10)+0.7Beta(17,4) | l1 and pointwise convergence: n=1024,4096,16384, 65536 | Improved |
| Use SURF instead of MATLAB fitdist Epanechnikov and Gaussian kernel estimators | SURF · 1994 US census salaries and sensorless drive data | test perplexity | Improved |
Closest papers by meaning
- TURF: Two-Factor, Universal, Robust, Fast Distribution Learning Algorithm (ICML 2022 · similarity 0.90)
- Unified Sample-Optimal Property Estimation in Near-Linear Time (NeurIPS 2019 · similarity 0.90)
- Optimal Algorithms for Augmented Testing of Discrete Distributions (NeurIPS 2024 · similarity 0.87)
- Optimal Hypothesis Selection in (Almost) Linear Time (NeurIPS 2024 · similarity 0.85)
- Nearly-Linear Time Private Hypothesis Selection with the Optimal Approximation Factor (NeurIPS 2025 · similarity 0.85)
- Sample-Adaptivity Tradeoff in On-Demand Sampling (NeurIPS 2025 · similarity 0.85)
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.
