The International Conference for High Performance Computing, Networking, Storage, and Analysis
The International Conference for High Performance Computing, Networking, Storage, and Analysis
Poster Type: Research Posters
Author: Anton Lebedev (STFC Hartree Centre), Won Kyung Lee (STFC Hartree Centre), Soumyadip Ghosh (IBM Thomas J. Watson Research Center), Olha I. Yaman (STFC Hartree Centre), Vassilis Kalantzis (IBM Thomas J. Watson Research Center), Yingdong Lu (IBM Thomas J. Watson Research Center), Tomasz Nowicki (IBM Thomas J. Watson Research Center), Shashanka Ubaru (IBM Thomas J. Watson Research Center), Lior Horesh (IBM Thomas J. Watson Research Center), Vassil Alexandrov (STFC Hartree Centre)
Supervisor:
Abstract: Large, sparse linear systems are pervasive in modern science and engineering, and Krylov subspace solvers are an established means of solving them. Yet convergence can be slow for ill-conditioned matrices, so practical deployments usually require preconditioners. Markov chain Monte Carlo (MCMC)-based inversion can generate such preconditioners and accelerate Krylov iterations, but its effectiveness depends on parameters whose optima vary across matrices; manual or grid search is costly. We present an AI-driven framework recommending MCMC parameters for a given linear system. A graph neural surrogate predicts preconditioning speed from A and MCMC parameters. A Bayesian acquisition function then chooses the parameter sets most likely to minimize iterations. On a previously unseen ill-conditioned system, the framework achieves better preconditioning with 50% of the search budget of conventional methods, yielding about a 10% reduction in iterations to convergence. These results suggest a route for incorporating MCMC-based preconditioners into large-scale systems.
Best Poster Finalist (BP): no
Poster: PDF
Poster Summary: PDF