welcome to the
Homepage of
Tobias Glasmachers

Table of Contents

Homepage of Tobias Glasmachers - Contact

Address

Building NB, Room 3/27
Institut für Neuroinformatik
Ruhr-Universität Bochum
Universitätsstr. 150
44780 Bochum, Germany

Telephone

+49-(0)234-32-25558

E-mail

Homepage of Tobias Glasmachers - Short CV

Homepage of Tobias Glasmachers - Teaching

Lectures and Seminars

In the winter term 2014/2015 I offer the following courses:

Previous Lectures and Seminars

Summer term 2014 "Machine Learning - Supervised Methods"
Summer term 2014 Seminar "Dimensionsreduktion und Lernen von Mannigfaltigkeiten"
Winter term 2013/14 "Evolutionäre Algorithmen"
Summer term 2012 "Machine Learning - Supervised Methods"
Winter term 2012/13 "Evolutionäre Algorithmen"
Summer term 2012 "Machine Learning - Supervised Methods"

Theses

I am always offering Bachelor and Master theses within the Angewandte Informatik program. Please contact me for details.

Homepage of Tobias Glasmachers - Research

My research is located in the area of machine learning, a modern branch of artificial intelligence research. This is an interdisciplinary research topic in between computer science, statistics, and optimization, with connections to neuroscience and applications in robotics, engineering, medicine, economics, and many more disciplines. Within this wide area I am focusing on two aspects: supervised learning, mostly with support vector machines, and evolutionary algorithms for search and optimization.

Supervised Learning

Supervised learning is a learning paradigm with endless (mostly technical) applications. A learning machine (algorithm) builds a predictive model from data provided in the form of input/output pairs. Primary examples are classification and regression problems. Support vector machines (SVMs) have advanced to a standard method in the field. On the one hand I am interested in the SVM training problem, which basically amounts to large scale quadratic programming. On the other hand I am trying to simplify SVM usage for non-experts by developing robust methods for automatic model selection. My research activities include both theoretical and practical aspects ranging from SVM optimization to experimental comparison studies and software development.

Evolutionary Algorithms

Evolutionary Algorithms (EAs) are a class of nature-inspired algorithms that mimic the process of Darwinian evolution. This process is resolved into the components inheritance, variation, and selection. It has been widely recognized that EAs are useful for search and optimization. Formally they can be understood as randomized direct search heuristics.

Homepage of Tobias Glasmachers - Research

Importantly, they are suitable for black-box optimization problems. I focus on evolution strategies, a class of optimization algorithms for continuous variables, and in multi-objective optimization. On the one hand I am interested in algorithm design and empirical studies. On the other hand the field offers fascinating theoretical challenges concerned with a deeper understanding of the mechanisms underlying direct randomized search. Last but not least I am interested in applications of EAs to relevant problems such as parameter tuning for learning machines or engineering problems.

Homepage of Tobias Glasmachers - Projects

Support-Vektor-Maschinen für extrem große Datenmengen

This two-years research project has started in November 2013. It is conducted in cooperation with the chair Computergestützte Statistik at the Technical University of Dortmund. It is funded by the Mercator Research Center Ruhr (MERCUR). A short project description can be found here.

Homepage of Tobias Glasmachers - Open Positions

Currently I don't have any open positions.

Homepage of Tobias Glasmachers - Software

Shark

I am an active developer of the Shark Machine Learning Library. Shark is an open-source, modular, and fast C++ library. A large share of my research code is either part of the library or based thereon. Check it out!

Shark is currently undergoing a major transition; actually it is more fair to speak of a complete rewrite. By now (2013) the work is mostly done, and we already have an alpha release of the brand new Shark 3. A few more design changes are underway that will make the library even faster and even more complete.

Adaptive Coordinate Frequencies Coordinate Descent

Coordinate descent with online adaptation of coordinate frequencies for fast training of linear models.

Hypervolume Maximization

Maximization of dominated hypervolume for multi-objective benchmark problems.

Homepage of Tobias Glasmachers - Presentations

Here you can find selected slides.