Senior Lecturer (Assistant Professor)
Dept. of Industrial Engineering and Management,
Ben Gurion University of the Negev, Israel.
P.O. 653, Beer Sheva, 84105
Mail: johnros <spammer decoy> bgu ac il
Distributed Machine Learning: Today’s BIG computing problems are solved, not with super-computers, but rather with thousands of smaller computers working together. This requires redesigning classical algorithms, originally designed for the serial computing age, and applying them to distributed computing environments. My research focuses on doing just that, for machine learning algorithms (and thus for other stochastic optimization problems).
On Brain Maps: I am interested in the methodological problems of brain imaging: How to generalize findings from a group study to a “typical brain”? How to analyze imaging genetic studies where the number of relations to study is in the Billions \((10^9)\)? How to account for the bias introduced by search-and-select methods such as region selection in social neuroscience (“Voodoo correlations”), and searchlight analysis in multivoxel pattern recognition?
High Dimensional Statistical Process Control: Modern day quality control does no longer consist of the variation of a single attribute over time. Production lines today have hundreds and thousands of sensors monitoring performance. It so happens, that the problem of quality control has applications in any monitoring problem. This include cyber security alerting systems, patient monitoring, internet of things (IOT), etc.
Statistical arguments and the philosophy of science: What makes a statistical finding “convincing”? Are all arguments (Bayesian, Frequentist, Fiducial,…) adequate for all problems?
I am actively recruiting PhD students and Post-docs with strong mathematical background to work on projects related to machine learning.
See my Google Scholar profile.
I like to publish my class notes. Here are the notes of some courses I give. They are not error free, so use at your own discretion and feel free to contact me for clarifications and errata.
- Quality Engineering, a.k.a. Industrial Statistics.
- R. A graduate course for learning how to analyze data with R. Hebrew videos now online!
- History and Basics: Hebrew, English.
- Basics continued: Hebrew,English.
- data.table: Hebrew. English.
- Exploratory Statistics: Hebrew, English1, English2.
- Linear Models: Hebrew, English.
- Generalized Linear Models: Hebrew, English.
- Linear Mixed Models: Hebrew, Ensligh.
- Multivariate Statistics: Hebrew, English.
- Supervised Machine Learning.
- Unsupervised Machine Learning:
- Plotting: Hebrew, English.
- Automated Reports:
- Sparse Linear Algebra: Hebrew, Ensligh1, English2.
- Memory Efficiency: Hebrew, English.
- Parallel Computing.
- Machine Learning and Data Mining.
- Dimesionality Reduction: a one class intro to dimensionality reduction.
Father (x3), beach-volleyball player, and data-hacker(*).
- Presently, a faculty member at the Dept. of Industrial Engineering and Management at the Ben Gurion University of the Negev, and also:
- PhD on statistical methods for neuroscience with Yoav Benjamini at Tel Aviv University.
(*) statistics + machine-learning + data science + programming + system administration.
If you are looking for a good data scientist in Israel, I am afraid you are not alone. Some places you may reach out to our graduates, and other data scientists:
My mail is at the head of this page. Alternatively, just follow the links at the bottom of the page.
I also moderate BGU’s computing-for-research listsrv. Join here (only accesible from within the BGU’s VPN).