Bio

I am a research scientist at Meta Platforms.

Previously, I was a PhD student in the LPNA and CS Department at UIUC, advised by Edgar Solomonik. Before that, I received my Master's degree from EECS Department at UC Berkeley and I finished my Undergraduate studies from Chu Kochen Honors College at Zhejiang University.

My research interests lie in the intersection of numerical algorithms, high performance computing, system and quantum simulation. In particular, I’m now focusing on developing efficient systems and numerical algorithms for tensor computations with applications in data analytics and quantum simulation. Previously, I worked on optimization and efficient numerical implementation of neural networks.

Here is my Curriculum Vitae.

• Towards Efficient Algorithms and Systems for Tensor Decompositions and Tensor Networks
  L. Ma
  Ph.D. Dissertation, University of Illinois Urbana-Champaign (2023)

  Paper Slides

• Tensor Rank and Other Multipartite Entanglement Measures of Graph States
  L. Schatzki, L. Ma, E. Solomonik and E. Chitambar
  Preprint: arXiv:2209.06320 (2022)

  Paper

• Cost-efficient Gaussian Tensor Network Embeddings for Tensor-structured Inputs
  L. Ma and E. Solomonik
  NeurIPS 2022, Preprint: arXiv:2205.13163 (2022)

  Paper

• Low Rank Approximation in Simulations of Quantum Algorithms
  L. Ma and C. Yang
  Journal of Computational Science, Preprint: arXiv:2104.11396 (2022)

  Paper Slides-SIAMCSE21

• Accelerating Alternating Least Squares for Tensor Decomposition by Pairwise Perturbation
  L. Ma and E. Solomonik
  Numerical Linear Algebra with Applications, Preprint: arXiv:1811.10573 (2022)

  Paper Code Slides-UCB-SCSeminar19 Slides-SIAMPP20 Poster-SIAMPP20

• Fast and Accurate Randomized Algorithms for Low-rank Tensor Decompositions
  L. Ma and E. Solomonik
  NeurIPS 2021, Preprint: arXiv:2104.01101 (2021)

  Paper Slides-SIAMLA21 Slides-Neurips21 Poster-Neurips21

• MLPruning: A Multilevel Structured Pruning Framework for Transformer-based Models
  Z. Yao, L. Ma, S. Shen, K. Keutzer, and M. W. Mahoney
  Technical Report, Preprint: arXiv:2105.14636 (2021)

  Paper

• Comparison of Accuracy and Scalability of Gauss-Newton and Alternating Least Squares for CP Decomposition
  N. Singh, L. Ma, H. Yang, and E. Solomonik
  SIAM Journal on Scientific Computing (SISC), Preprint: arXiv:1910.12331 (2021)

  Paper Code

• Efficient parallel CP decomposition with pairwise perturbation and multi-sweep dimension tree
  L. Ma and E. Solomonik
  International Parallel and Distributed Processing Symposium (IPDPS), Preprint: arXiv:2010.12056 (2021)

  Paper Slides-IPDPS21 Code

• AutoHOOT: Automatic High-Order Optimization for Tensors
  L. Ma, J. Ye, and E. Solomonik
  International Conference on Parallel Architectures and Compilation Techniques (PACT'20), Preprint: arXiv:2005.04540 (2020)

  Paper Code Slides-PACT20 Video

• Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT
  S. Shen, Z. Dong, J. Ye, L. Ma, Z. Yao, A. Gholami, M. W. Mahoney, and K. Keutzer,
  AAAI 2020, Preprint: arXiv:1909.05840 (2020)

  Paper Poster-AAAI20

• Inefficiency of K-FAC for Large Batch Size Training
  L. Ma, G. Montague, J. Ye, Z. Yao, A. Gholami, K. Keutzer, and M. W. Mahoney,
  AAAI 2020, Preprint: arXiv:1903.06237 (2020)

  Paper Poster-AAAI20

• A Multiscale Model for Electrochemical Reactions in LSCF Based Solid Oxide Cells
  L. Ma, P. Priya, and N. R. Aluru,
  Journal of the Electrochemical Society (2018)

  Paper Slides-USNCCM17

• A Multiscale Model for the Oxide Ion Conducting and Proton Conducting Solid Oxide Cells
  L. Ma
  M.S. Thesis, University of Illinois Urbana-Champaign (2018)

  Paper Slides

• An empirical study of neural ordinal differential equations

  Report Poster

• Convex analysis of Nonnegative Matrix Factorization

  Report

• lma16@illinois.edu
• (217) 979-7114