Uncertainty-Calibrated Collaborative Filtering via Heteroscedastic Bayesian Neural Networks

Short summary

BayesNCF-v2, a novel Bayesian neural collaborative filtering framework, jointly learns both aleatoric and epistemic uncertainty, outperforming standard NCF by reducing ECE by 25-73% and NLL by 7-20% across three datasets.

AI-generated from the title and abstract; the full text is not read.