Paper 1: [Best Paper Award] PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback



Yuncong Yu, Dylan Kruyff, Jiao Jiao, Tim Becker, Michael Behrisch

 View presentation:2022-10-16T14:41:00ZGMT-0600Change your timezone on the schedule page
2022-10-16T14:41:00Z

Exemplar figure, described by caption below — PSEUDo is an efficient, adaptive and interpretable tool for visual pattern retrieval in multivariate time series. It makes locality-sensitive hashing trainable while maintaining its efficiency and promoting interpretability. It introduced a relevance feedback mechanism to capture subjective similarity through feature selection. PSEUDo is particularly efficient for very high-dimensional time series and in cases where initial labels are meager, and the promptness of the outcome counts.

The live footage of the talk, including the Q&A, can be viewed on the session page, VDS: Opening / Keynote 1 + Paper Session 1.

Abstract