Cameron Musco

Stata Center | 32 Vassar Street, Cambridge, MA 02139 | Office 32-G670 | cnmusco at mit dot edu
my photo

I am a fifth year Ph.D. student (graduating spring 2018) in the Theory of Computation Group at MIT.

I am advised by Nancy Lynch and supported by an NSF Graduate Fellowship. I study algorithms, focusing on applications in data science and machine learning. I often work on randomized methods and algorithms that adapt to streaming and distributed computation. I am also interested in understanding randomized computation and algorithmic robustness by studying computational processes in biological systems.

Before MIT, I studied Computer Science and Applied Mathematics at Yale University and worked as a software developer on the Data Team at Redfin.

Here are my Google Scholar profile, CV, and GitHub.

The website for our reading group on distributed computation in biological systems can be found here.

Publications

Algorithms and Large-Scale Linear Algebra

Spectrum Approximation Beyond Fast Matrix Multiplication: Algorithms and Hardness
Cameron Musco, Praneeth Netrapalli, Aaron Sidford, Shashanka Ubaru, and David P. Woodruff
Innovations in Theoretical Computer Science (ITCS) 2018.

Stability of the Lanczos Method for Matrix Function Approximation
Cameron Musco, Christopher Musco, and Aaron Sidford
ACM-SIAM Symposium on Discrete Algorithms (SODA) 2018.
Matlab code for matrix function approximation (see lanczos.m).

Recursive Sampling for the Nyström Method
Cameron Musco and Christopher Musco
Conference on Neural Information Processing Systems (NIPS) 2017.
Matlab code.

Is Input Sparsity Time Possible for Kernel Low-Rank Approximation?
Cameron Musco and David P. Woodruff
Conference on Neural Information Processing Systems (NIPS) 2017.

Sublinear Time Low-Rank Approximation of Positive Semidefinite Matrices
Cameron Musco and David P. Woodruff
IEEE Symposium on Foundations of Computer Science (FOCS) 2017.
Slides from my talk at FOCS. Extended slides slides for hour long talk.

Random Fourier Features for Kernel Ridge Regression: Approximation Bounds and Statistical Guarantees
Haim Avron, Michael Kapralov, Cameron Musco, Christopher Musco, Ameya Velingker and Amir Zandieh
International Conference on Machine Learning (ICML) 2017.
Slides from my talk at ICML.

Input Sparsity Time Low-Rank Approximation via Ridge Leverage Score Sampling
Michael B. Cohen, Cameron Musco, and Christopher Musco
ACM-SIAM Symposium on Discrete Algorithms (SODA) 2017.
Slides from my talk at SODA. Chris's slides from his talk at University of Utah.

Online Row Sampling
Michael B. Cohen, Cameron Musco, and Jakub Pachocki
International Workshop on Approximation Algorithms for Combinatorial Optimization Problems (APPROX) 2016.
Invited to special issue of Theory of Computing.

Principal Component Projection Without Principal Component Analysis
Roy Frostig, Cameron Musco, Christopher Musco, and Aaron Sidford
International Conference on Machine Learning (ICML) 2016.
Matlab code.

Faster Eigenvector Computation via Shift-and-Invert Preconditioning
Daniel Garber, Elad Hazan, Chi Jin, Sham M. Kakade, Cameron Musco, Praneeth Netrapalli, and Aaron Sidford
International Conference on Machine Learning (ICML) 2016.

Randomized Block Krylov Methods for Stronger and Faster Approximate Singular Value Decomposition
Cameron Musco and Christopher Musco
Conference on Neural Information Processing Systems (NIPS) 2015.
Selected for Oral Presentation (1 of 15 out of 403 papers).
Slides from my talk at NIPS. Matlab code.

Dimensionality Reduction for k-Means Clustering and Low Rank Approximation
Michael B. Cohen, Sam Elder, Cameron Musco, Christopher Musco, and Madalina Persu
ACM Symposium on Theory of Computing (STOC) 2015.
Slides from my talk at MIT's Algorithms and Complexity Seminar.
My Master's Thesis containing empirical evaluation along with a guide to implementation.

Uniform Sampling for Matrix Approximation
Michael B. Cohen, Yin Tat Lee, Cameron Musco, Christopher Musco, Richard Peng, and Aaron Sidford
Innovations in Theoretical Computer Science (ITCS) 2015.
Slides from my talk at MIT's Algorithms and Complexity Seminar.

Single Pass Spectral Sparsification in Dynamic Streams
Michael Kapralov, Yin Tat Lee, Cameron Musco, Christopher Musco and Aaron Sidford
IEEE Symposium on Foundations of Computer Science (FOCS) 2014.
Appeared in special issue of SIAM Journal on Computing.
Chris's Slides from his talks at FOCS and the Harvard TOC Seminar.

Biological Computation

Neuro-RAM Unit with Applications to Similarity Testing and Compression in Spiking Neural Networks
Nancy Lynch, Cameron Musco, and Merav Parter
International Symposium on Distributed Computing (DISC) 2017.

Spiking Neural Networks: An Algorithmic Perspective
Nancy Lynch, Cameron Musco, and Merav Parter
Presentation at Workshop on Biological Distributed Algorithms (BDA) 2017.
Slides from my talk at BDA.

New Perspectives on Algorithmic Robustness Inspired by Ant Colony House-Hunting
Tsvetomira Radeva, Cameron Musco, and Nancy Lynch
Presentation at Workshop on Biological Distributed Algorithms (BDA) 2017.

Computational Tradeoffs in Biological Neural Networks: Self-Stabilizing Winner-Take-All Networks
Nancy Lynch, Cameron Musco, and Merav Parter
Innovations in Theoretical Computer Science (ITCS) 2017.

Ant-Inspired Density Estimation via Random Walks
Cameron Musco, Hsin-Hao Su, and Nancy Lynch
Proceedings of the National Academy of Sciences (PNAS), 2017.
Full paper also available on arXiv. An extended abstract initially appeared in PODC 2016.

Distributed House-Hunting in Ant Colonies
Mohsen Ghaffari, Cameron Musco, Tsvetomira Radeva, and Nancy Lynch
ACM Symposium on Principles of Distributed Computing (PODC) 2015.


Other Writing

Here are a few writeups, notes, and talks.

VC dimension in (neural) circuit complexity, outline of a talk on the basics of VC dimension and how it can be used to give circuit size lower bounds for certain functions.

Fast Low-Rank Approximation and PCA Beyond Sketching, slides for a talk I gave at Mining Massive Datasets (MMDS 2016) on new techniques for large scale low-rank approximation.

Chebyshev Polynomials in TCS and Algorithm Design, outline of a talk I gave at the MIT Theory student retreat on the many applications of Chebyshev polynomials to upper and lower bounds in Theoretical Computer Science.

Subspace Scores for Feature Selection in Computer Vision, final project report where we test leverage score based sampling algorithms for k-means/PCA dimension reduction and general feature selection.

Applications of Linear Sketching to Distributed Computing, slides for a talk I gave at our Theory of Distributed Systems seminar. High level overview of linear sketching, my work on k-means clustering and spectral sparsification, and applications to distributed data analysis.

Graph Sparsification and Dimensionality Reduction, final project for Jelani Nelson's Algorithms for Big Data class.

Linear Regression and Pseudoinverse Cheatsheet, since there are a lot of ways to explain the pseudoinverse.

Big-O and Asymptotic Notation Cheatsheet, just in case.

Fast Approximation of Maximum Flow using Electrical Flows, undergraduate Applied Mathematics senior project report. I was fortunate to be advised by Dan Spielman for both my senior projects. They were my first introduction to research in computer science and the reason I decided to go to graduate school.

Graph Construction Through Laplacian Function Optimization, Computer Science senior project report.


Side Projects

I've worked on a lot of projects, some more serious than others.

I built this Rap Collaboration Graph, which gives a visualization of musical collaborations in hip hop. Unfortunately, colors are only rendering in Safari right now, it looks fuzzy on retina displays, and the data is about a year out of date. But I promise I'll get to it.

I had a lot of fun helping build the first version of a site that sold buffalo chicken sandwiches. It's gotten a major facelift and evolved into Crunchbutton, but here is a screenshot of the original One Button Wenzel in its glory.

My friend Charlie and I once built an AI to play Transport Tycoon Deluxe. Here is a poster describing the project.

In college I also had a lot of fun working on Yale's Formula Hydrid Racecar Team.

I love to ski, and a long time ago, my brother Chris and I used to build custom skis. Our first pair had maple/poplar cores, varnished wood sidewalls, and flex comparable to a pair of 2x4s. Our second pair has more precisely shaped pine cores, an improved top sheet, and a much smoother flex. The tips delaminated once, but we riveted them together and still ski on them today! Here's an old forum post on SkiBuilders.com with more pictures of our setup.