@article{multiheadattentionasensemblenadarayawats, title = {Multi-Head Attention as Ensemble Nadaraya-Watson Estimation: Variance Reduction, Decorrelation, and Optimal Head Diversity}, author = {Ernest Fokoué}, year = {2026}, eprint = {2605.20271}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2605.20271}, }