Publications

Published/Accepted

  1. Projective Splitting with Forward Steps, P. Johnstone and Jonathan Eckstein, arXiv:1803.07043, March 2018, (accepted Mathematical Programming). [code], [youtube], [slides].

  2. Single-Forward-Step Projective Splitting: Exploiting Cocoercivity, P. Johnstone and Jonathan Eckstein, arXiv:1902.09025, February 2019 (accepted Computational Optimization and Applications (COAP), Oct 2020).

  3. Convergence Rates for Projective Splitting, P. Johnstone and Jonathan Eckstein, SIAM Journal on Optimization 29(3), 1931–1957, July 2019, DOI 10.1137/18M1203523.

  4. Faster Subgradient Methods for Functions with Hölderian Growth, P. Johnstone and Pierre Moulin, Mathematical Programming (available online), January 2019, DOI 10.1007/s10107-018-01361-0.

  5. Local and Global Convergence of a General Inertial Proximal Splitting Scheme for Minimizing Composite Functions, P. Johnstone and Pierre Moulin, Computational Optimization and Applications (COAP), 67(2) pp. 259-292 (2017), DOI 10.1007/s10589-017-9896-7. (This manuscript is a thoroughly revised and improved version of an earlier preprint from March 2015 titled A Lyapunov analysis of FISTA with local linear convergence for sparse optimization)

  6. Projective Splitting with Forward Steps only Requires Continuity, P. Johnstone and Jonathan Eckstein, Optimization Letters (2019) available online, DOI 10.1007/s11590-019-01509-7

Preprints

  1. Stochastic Projective Splitting: Solving Saddle-Point Problems with Multiple Regularizers, P. Johnstone, Jonathan Eckstein, Thomas Flynn, and Shinjae Yoo, arXiv preprint arXiv:2106:1306 June 2021.

Refereed Conference Papers

  1. Convergence Rates of Inertial Splitting Schemes for Nonconvex Optimization, P. Johnstone and Pierre Moulin, Proc. ICASSP, March 2017, New Orleans USA.

  2. Strong Large Deviations for Rao Test Score and GLRT in Exponential Families, Pierre Moulin and P. Johnstone, Proc. ISIT, June 2015, Hong Kong.

  3. Convergence of an Inertial Proximal Method for L1-Regularized Least-Squares, P. Johnstone and Pierre Moulin, Proc. ICASSP, April, 2015, Brisbane, Australia.

  4. Kullback-Leibler Divergence and the Central Limit Theorem, Pierre Moulin and P. Johnstone, Proc. ITA, San Diego, CA, Feb. 2014.

Theses

  1. Accelerated First-Order Optimization Methods Using Inertia And Error Bounds, P. Johnstone, PhD Thesis, UIUC, May 2017.

  2. Inertial Iterative Thresholding with Applications to Sparse and Low-Rank Signal Recovery, P. Johnstone, Masters Thesis, UIUC, August 2014.

  3. Point Process Modelling and Spectral Analysis in Neural Coding, P. Johnstone, Bachelors Honors Thesis, UNSW, October 2009