Marthinus Christoffel (Christo) du Plessis
Assistant Professor, University of Tokyo
Education
Bachelor, Bachelor (Hons), cum laude
- Computer Engineering
- University of Pretoria
- Best software engineering project
Master of Engineering, cum laude
- Computer Engineering
- University of Pretoria
- Thesis title: “Non-stationary signal classification for radar transmitter identification”
- Supervisor: prof. J.C. Olivier
Doctor of Engineering
- Computer Science (Machine learning)
- Thesis title: “Learning under Class-Balance Change: Distribution Matching via Direct Divergence Estimation”
- Supervisor: prof. M. Sugiyama
Work Experience
The University of Tokyo, Assistant Professor
- 2014.4 – current
- Research in machine learning
- Implementation of algorithms in C/C++ and Python
The University of Tokyo, Post-doctoral researcher
Tokyo Institute of Technology, Post-doctoral researcher
- 2013.4 – 2013.9 (Tokyo Inst. Tech), 2013.10 – 2013.3 (UTokyo)
- Research in machine learning (several papers at ICML/NIPS)
- Implementation of algorithms in C/C++ and Python
Skills and competencies
- Strong mathematics and statistics foundation
- Data visualization experience with a variety of packages (MATLAB, Matplotlib toolkit)
- Skilled in data manipulation
- Excellent knowledge of machine learning methods for high-dimensional dataset analysis (dimensionality reduction, feature selection, sparsity inducing regularization, etc…)
- Experience in application of high-volume learning methods and development of new methods (see my online learning publications)
In-depth knowledge of a wide variety of data analysis methods:- Cluster analysis (k-means, spectral clustering, etc…)
- Classification (SVM, neural networks, logistic regression, boosted decision trees, etc…)
- Regression analysis (least-squares, Huber loss, SVR, etc…)
- Anomaly/outlier detection (1-class SVM, SVDD, PU learning)
- Programming languages: Python (numpy/scipy, scikit), Matlab/Octave, C++
- Databases: SQL (MySQL)