We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: Can a single neuron learn quantiles?

Abstract: A novel non-parametric quantile estimation method for continuous random variables is introduced, based on a minimal neural network architecture consisting of a single unit. Its advantage over estimations from ranking the order statistics is shown, specifically for small sample size. In a regression context, the method can be used to quantify predictive uncertainty under the split conformal prediction setting, where prediction intervals are estimated from the residuals of a pre-trained model on a held-out validation set to quantify the uncertainty in future predictions. Benchmarking experiments demonstrate that the method is competitive in quality and coverage with state-of-the-art solutions, with the added benefit of being more computationally efficient.
Comments: 8 pages (main) + 9 pages (supplementary material). Source code provided
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2106.03702 [stat.ML]
  (or arXiv:2106.03702v2 [stat.ML] for this version)

Submission history

From: Edgardo Solano Carrillo [view email]
[v1] Mon, 7 Jun 2021 15:12:47 GMT (233kb,D)
[v2] Tue, 3 Aug 2021 11:20:02 GMT (242kb,D)

Link back to: arXiv, form interface, contact.