Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Heteroscedastic Neural Networks for Path Loss Prediction with Link-Specific Uncertainty

About

Traditional and modern machine learning-based path loss models typically assume a constant prediction variance. We propose a neural network that jointly predicts the mean and link-specific variance by minimizing a Gaussian negative log-likelihood, enabling heteroscedastic uncertainty estimates. We compare shared, partially shared, and independent-parameter architectures using accuracy, calibration, and sharpness metrics on blind test sets from large public RF drive-test datasets. The shared-parameter architecture performs best, achieving an RMSE of 7.4 dB, 95.1 percent coverage for 95 percent prediction intervals, and a mean interval width of 29.6 dB. These uncertainty estimates further support link-specific coverage margins, improve RF planning and interference analyses, and provide effective self-diagnostics of model weaknesses.

Jonathan Ethier• 2025

Related benchmarks

TaskDatasetResultRank
Path Loss PredictionLOGO-CV Blind 6 holdouts (test)
RMSE Mean (dB)7.4
4
Path Loss PredictionLondon-only (test)
RMSE (dB)7.1
4
Showing 2 of 2 rows

Other info

Follow for update