Avatar

Michiel Straat

PhD Student Intelligent Systems

University of Groningen

Biography

I am a PhD student in the Intelligent Systems group at the Bernoulli Institute for Mathematics, Computer Science and Artificial Intelligence. In the project, we use techniques from statistical physics to analyze the learning behavior of machine learning models and get more insight into practically relevant phenomena, which can be used to improve machine learning algorithms.

Interests

  • Machine Learning
  • Computational Intelligence
  • Statistical Physics
  • Data Science

Education

  • PhD in Computing Science, 2018-2022

    University of Groningen

  • MSc. in Computing Science, 2016-2018

    University of Groningen, MSc. thesis

  • BSc. in Computing Science, 2013-2016

    University of Groningen, BSc. thesis

Recent Publications

Quickly discover relevant content by filtering publications.

Preprint: Hidden Unit Specialization in Layered Neural Networks: ReLU vs. Sigmoidal Activation

Our systematic comparison of networks with ReLU and sigmoidal units in model situations reveals surprising differences in their …

On-line learning dynamics of ReLU neural networks using statistical physics techniques

We introduce exact macroscopic on-line learning dynamics of two-layer neural networks with ReLU units in the form of a system of …

Learning Vector Quantization and relevances in complex coefficient space

In this contribution, we consider the classification of time series and similar functional data which can be represented in complex …

Statistical Mechanics of On-Line Learning Under Concept Drift

We introduce a modeling framework for the investigation of on-line machine learning processes in non-stationary environments. We …

Prototypes and matrix relevance learning in complex Fourier space

In this contribution, we consider the classification of time-series and similar functional data which can be represented in complex …

Segmentation of blood vessels in retinal fundus images

In recent years, several automatic segmentation methods have been proposed for blood vessels in retinal fundus images, ranging from …

Recent & Upcoming Talks

Towards a statistical physics analysis of multilayer ReLU neural networks

On-line learning dynamics of ReLU neural networks using statistical physics techniques

Online Learning Dynamics of ReLU Neural Networks

The rectifier activation function (Rectified Linear Unit: ReLU) has become popular in deep learning applications, mostly because the …

On-line learning in neural networks with ReLU activations

The rectifier activation function (Rectified Linear Unit: ReLU) has become popular in deep learning applications, mostly because the …

Prototypes and matrix relevance learning in complex Fourier space

Segmentation of blood vessels in retinal fundus images

The inspection of the blood vessel tree in the fundus, which is the interior surface of the eye opposite to the lens, is important in …

Teaching

 
 
 
 
 

Head of Teaching Assistants

Advanced Algorithms and Data Structures

November 2019 – March 2020 University of Groningen
Advanced course in algorithms and data structures. Students bring the theory into practice by solving algorithmic challenges every week that are prepared and graded by the teaching assistants. More information on the course can be found here.
 
 
 
 
 

Teaching Assistant (2x)

Information Security

September 2016 – December 2017 University of Groningen
Course on information security. Students learn the fundamental cryptographic algorithms and their properties that are at the core of security of modern systems. Understanding of the methods is increased by exercises in which students implement the algorithms themselves. More information on the course can be found here.

Contact