Thesis Information
Title: Universality of Random Matrices and Random Graphs
Adviser: Van Vu
Institution: Yale University
Graduation Date: May 2017
Contact Information
Email Candidate
Candidate Website
Candidate Bio:
I am an NSF postdoctoral fellow at Harvard University with joint appointments in the School of Engineering and Applied Sciences and the Center for Mathematical Sciences and Applications. I graduated from Yale University in 2017 under the supervision of Van Vu. Before that, I received my undergraduate degree from Harvey Mudd College.
Teaching Aims:
I have seven years of teaching experience and have won two competitive teaching awards. I recently designed and taught a graduate course on high-dimensional probability at Harvard. In the classroom, I try to achieve a balance between intuition and rigor. I am also committed to promoting participation in mathematics and computer science from traditionally underrepresented groups.
Paper 1:
Dictionary learning with few samples and matrix concentration, K. Luh and Van Vu, IEEE Transactions on Information Theory, 2016
Link to PDFPaper 2:
Sparse random matrices have simple spectrum, K. Luh and Van Vu, to appear in Annales de l'Institut Henri Poincare, 2019
Link to PDFPaper 3:
An improved lower bound for sparse reconstruction from subsampled Hadamard matrices, J. Blasiok, P. Lopatto, K. Luh, J. Marcinek, S. Rao, to appear in FOCS 2019
Link to PDFKeywords: Random matrices, high-dimensional algorithms, machine learning, compressed sensing