• Facebook
  • LinkedIn
  • Twitter

Dr. Hao Ji

Institution: California State Polytechnic University, Pomona

Department: Computer Science

Proposed research ideas: 

Large-scale linear solvers play an important role in various practical applications, ranging from data analysis to large-scale scientific computing. Deflated block Krylov subspace methods exhibit many attractive properties when solving large linear systems with multiple right-hand sides, including accelerated convergence rate, reduced data accesses, and intrinsic parallelism. Our past effort is analyzing and developing breakdown-free block Conjugate Gradient algorithms with deflation to address the potential rank deficiency and gain convergence acceleration in solving least-squares problems. We expect research collaboration on designing fast large-scale linear solvers, i.e., accelerating block solvers using deflation techniques and designing scalable implementations of deflated block Krylov subspace algorithms over modern parallel/distributed computing platforms. We also hope to collaborate researchers from different disciplines to apply deflated block linear solvers to real-world problems.

Motivation:

My primary research interest lies in the field of large-scale linear algebra. This program would provide me with an important opportunity to introduce my research work to leading experts in this field during the workshop. Their comments and suggestions would greatly help me find potential collaborations and develop competitive research proposals, especially for the fundamental research on linear solver field. As a new assistant professor, I also expect to explore potential early career research opportunity for building a successful research career.