CS Forum: Sahely Bhadra, Indian Institute of Technology Palakkad "Warping Resilient Time Series Embeddings"
Warping Resilient Time Series Embeddings
Prof. Sahely Bhadra
Department of Computer Science and Engineering
Indian Institute of Technology (IIT) Palakkad
Time series are ubiquitous in real world problems and computing distance between two time series is often required in several learning tasks. Computing similarity between time series by ignoring variations in speed or warping is often encountered and dynamic time warping (DTW) is the state of the art. However DTW is not applicable in algorithms which require kernel or vectors. In this paper, we propose a mechanism named WaRTEm to generate vector embeddings of time series such that distance measures in the embedding space exhibit resilience to warping. Therefore, WaRTEm is more widely applicable than DTW. WaRTEm is based on a twin auto-encoder architecture and a training strategy involving warping operators for generating warping resilient embeddings for time series datasets. We evaluate the performance of WaRTEm and observed more than $20\%$ improvement over DTW in multiple real-world datasets.
Sahely Bhadra is assistant professor in Computer Science and Engineering of Indian Institute of Technology (IIT), Palakkad. Her research interest is Machine Learning and optimization for multi-view and noisy data. She has received her PhD from IISc,Bangalore. Before joining IIT Palakkad, Sahely worked as postdoctoral researcher in Aalto University, Finland and MPI-informatics, Germany.
Prof. Bhadra is visiting Prof. Juho Rousu during 25.6-17.7. Email the speaker ([email protected]) if you wish to have a one-on-one discussion.