Parthasarathy directs Ohio State’s Data Mining Research Laboratory, which is a part of the High End Systems Group and affiliated with the Laboratory for Artificial Intelligence Research. He was among the pioneers to enhance the locality of various frequent pattern mining algorithms while facilitating incremental, compressed and parallel analysis. He co-led the development of the ECLAT family of algorithms, popularized in leading textbooks in the field. He has designed award-winning novel architecture-conscious data structures and tiling strategies to enhance chip utilization and out-of-core performance on single node and cluster systems.
Parthasarathy also pioneered fundamental ideas for analyzing evolving networks, including his award-winning event-based framework and a simple probabilistic framework. He has designed groundbreaking algorithms for extracting robust modular and community structure from networks. He has developed new methods for graph representation learning that explicitly account for outliers with downstream applications for flood mapping from satellite imagery and biological network analysis.