Alexandra Bernadotte

This paper introduces a geometric framework for time series classification using delay embeddings and topology, achieving state-of-the-art performance on EEG data without feature engineering.