Homepage of Jun Liu


       I am a principal data scientist and co-founder for Infinia ML, located in Durham, North Carolina.
       I was a principal research statistician developer for SAS Institute Inc. located at Cary, North Carolina. My focus was on model selection, sparse learning, big data, online learning and so on.
       I was a research scientist for Siemens Corporate Research located at Princeton, New Jersey. My focus was on parallel MRI reconstruction and sparse learning. My developed algorithms, as demonstrated in 9+ issued US patents, run very well on hundreds of trials at the MR scanners achieving impressive image quality. A department (Siemens CT RTC ICV) award was given to this work.
       I worked at Arizona State University as a postdoc and then a research scientist in the lab of Dr. Jieping Ye. During the three years' work at Dr. Ye's lab, I mainly focused on large-scale sparse learning, and developed the SLEP package for feature selection in the large-scale scenario.
       I obtained my Ph.D. in computer science from Nanjing University of Aeronautics and Astronautics (NUAA), supervised by Prof. Songcan Chen. My Ph.D thesis "Research on Subspace Representation of Face Images" (in Chinese) is on Feature Extraction and Face Recognition. My Ph.D. thesis was awarded the Excellent PhD Thesis of Jiangsu Province, P. R. China, in 2009.
       I was born in a small village of the coastal city NanTong, P. R. China. 
       I have a happy family, with two lovely sons. My older son is helping us dealing with COVID-19 via Printed Reality.
       I can be reached via FirstNameLastName.NT@Gmail.com (FirstName=Jun, LastName=Liu).

Research Topics

  • Large-Scale Sparse Learning: learning a sparse and interpretable representation of the high-dimensional data, under different structure assumptions

  • Feature Extraction: conducting dimensionality reduction in the unsupervised (e.g., Principal Component Analysis) and the supervised (e.g., Linear Discriminant Analysis) manner

  • Face Recognition: performing face recognition under various facial variations