Research

Research Interests:

Spectral Methods for Machine Learning & Scientific Computation

• • • Spectral sparsification (coarsening) of undirected graphs (DAC'16, DAC'18, DAC'19, TCAD'20, software ICCAD'20, DAC'24, software), directed graphs (TKDD'24), and hypergraphs (ICCAD'21, ICCAD'22, software).

    • Spectral frameworks for stability assessment/enhancement of machine learning models (ICML'21, software) and graph neural networks (GNNs) (LoG'22, software). 

    • Spectral algorithms for data clustering (BMVC'22) and visualization (CG'20), multilevel graph partitioning (WSDM'21) and embedding (ICLR'20, software), solving SDD matrices (ICCAD'17), and graph topology learning from data (DAC'21, WSDM'22, TCAD'22, software).

    • Spectral methods for faster training of physics-informed neural networks (PINNs) for solving large parametric partial differential equations (DAC'24, software).

Scalable Methods for VLSI Design & Computer-aided Design

• • • Scalable SPICE and harmonic balance (HB) algorithms for IC simulations (TCAD'15a, TCAD'15b, ICCAD'13, ICCAD'12, DAC'12).

    • Fast algorithms for on-chip power grid modeling and simulation (TCAD'21, ICCAD'14, TVLSI'14, DATE'13, TODAES'11, ICCAD'11a, ICCAD'11b).

    • Robust design, optimization and verification of ICs  (TODAES'22, DATE'20, DAC'17, TVLSI'13, DAC'13a, DAC'10b, DAC'10c, TCAD'11b, TVLSI'10, ICCAD'06a).

    • Statistical VLSI modeling and analysis (TCAD'09, TVLSI'09, IET'09, TCAD'08, DAC'07a, DAC'07b, ICCAD'07b, ICCAD'07a, TVLSI'07, ICCAD'06b).

Heterogeneous Parallel Numerical & CAD Algorithms

• • • Hybrid multigrid solver on GPUs for power grid analysis (TCAD'11a, TVLSI'11, DAC'10a, ICCAD'08).

    • Fast multipole method (FMM) on GPU for capacitance extraction (DAC'11).

    • SPICE and harmonic balance (HB) simulators on GPU (ICCAD'16, DAC'15, DAC'13b).

    • Fast thermal simulator on GPU for 3D-ICs (TVLSI'13a, ICCAD'11b).

Research Grants/Gifts:

• (Sole PI) SHF: Small: High-Performance Incremental Spectral Algorithms for Efficient Modeling and Simulation of Large-Scale Integrated Circuits, Division of Computing and Communication Foundations, 2024-2027, The National Science Foundation. (Award Number: 2417619) 

• (Lead PI) SHF: Medium: Co-optimizing Spectral Algorithms and Systems for High-Performance Graph Learning, Division of Computing and Communication Foundations, 2022-2026, The National Science Foundation. (Award Number: 2212370)

• (Sole PI) SHF: Small: Learning Circuit Networks from Measurements, Division of Computing and Communication Foundations, 2022-2025, The National Science Foundation. (Award Number: 2205572) 

• (Sole PI) SHF: Small: Spectral Reduction of Large Graph and Circuits Networks, Division of Computing and Communication Foundations, 2019-2023, The National Science Foundation. (Award Number: 1909105, 2021309) 

• (Sole PI) SHF: Small: Scalable Spectral Sparsification of Graph Laplacians and Integrated Circuits, Division of Computing and Communication Foundations, 2016-2021, The National Science Foundation. (Award Number: 1318694, 2011412) 

• (Sole PI) CAREER: Leveraging Heterogeneous Manycore Systems for Scalable Modeling, Simulation and Verification of Nanoscale Integrated Circuits, Division of Computing and Communication Foundations, 2014-2022, The National Science Foundation. (Award Number: 2041519, 1350206) 

• (Sole PI) SHF: Small: Graph Sparsification Approach to Scalable Parallel SPICE-Accurate Simulation of Post-layout Integrated Circuits, Division of Computing and Communication Foundations, 2013-2017, The National Science Foundation. (Award Number: 1318694)

• (Sole PI) Spectral Methods for Scalable Integrated Circuit Partitioning and Analysis, Keysight Technologies

• (Sole PI) Consulting Support, NXP Semiconductors

• (Sole PI) Consulting Support, Cadence Design Systems

• (Sole PI) Hardware donations, XILINX, inc

• (Sole PI) GPU Research Center, NVIDIA Corporation

• (Sole PI) Leveraging Heterogeneous Manycore Systems for Scalable Modeling, Analysis and Verification of Nanoscale VLSI Systems, Research Excellence Fund, MTU

• (Sole PI) Verification of Large Power Delivery Networks, Intel Corporation

Honors & Awards:

• Institute of Computing and Cybersystems (ICC) Achievement Awards, MTU, 2017

• Top 10% faculty in instructor rating, MTU, 2016

• NSF CAREER Award, 2014

• Best Paper Award (the sole winner selected from 747 submitted papers), 50th IEEE/ACM Design Automation Conference (DAC'13)

• Best Paper Award Nomination, 2008 IEEE/ACM International Conference on Computer-Aided Design, (ICCAD'08)

• Best Paper Award Nomination, 2006 IEEE/ACM International Conference on Computer-Aided Design, (ICCAD'06)