Reproduce “Confidence-Nets: A Step Towards better Prediction Intervals for regression Neural Networks on small datasets” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Mohamedelmujtaba Altayeb, Abdelrahman M. Elamin, Hozaifa Ahmed, Eithar Elfatih Elfadil Ibrahim, Omer Haydar, Saba Abdulaziz, Najlaa H. M. Mohamed
Confidence-Nets adds an error-estimation model to a neural network, producing a corrected prediction and an interval intended to reflect expected variation. It trains XGBoost on prediction errors and widens the margin for unfamiliar inputs. On Concrete Compressive Strength, inclusion was 83% versus 0% for the base ANN at 90% training; this is not accuracy, and the smallest datasets performed worst.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.