Skip to ContentSkip to Navigation
Research Open Science Open Research Award

The Shallow and the Deep - A biased introduction to neural networks and old school machine learning.

Michael Biehl (Faculty of Science and Engineering)
Michael Biehl
Michael Biehl

Open Research objectives/practices

  • Openly sharing self-created or adapted educational materials using open licenses to enable and encourage reuse and customisation by others.
  • Using open materials to decrease the cost of education and/or increase the sustainability of (public) investments in education.

Introduction

I wrote an open textbook entitled The Shallow and the Deep - A biased introduction to neural networks and old school machine learning, which has been published in September 2023 by University of Groningen Press (open access and print-on-demand). Thankfully, the publication was supported by the Open Access Textbook Fund of the University of Groningen - which has unfortunately run out in the meantime. The textbook comprises full text lecture notes corresponding to my course Neural Networks and Computational Intelligence, given in the MSc programme Computing Science. Beyond its use as course material it is meant as a general introduction to machine learning to a broader audience. It also provides references and links to related materials and resources.

The recent popularity (or hype) of machine learning and artificial intelligence has triggered a lot of interest in the topic. At the same time, however, unrealistic expectations and the frequent uncritical and uninformed use of ready-made software tools has its risks.

In a sense, the goal of the book is to „de-mistify“ machine learning by providing a solid conceptual, theoretical and mathematical background. Consequently, the focus is on foundations and basic concepts, while the most recent developments and latest trends are not touched upon or treated only very briefly.

Motivation

The starting point for these notes was the desire to provide more comprehensive material than the lecture slides tot he students oft he MSc level course Neural Networks (later re-named to Neural Networks and Computational Intelligence). Through the years I had compiled a growing set of full text lecture notes, provided freely, chapter by chapter, to the students via the course website. Later I also made them available in the PURE system. Eventually, UG Press contacted me and suggested to publish the notes as an open access textbook (online and print-on-demand). Given the availability for fincancial support by the Open Textbook Fund, I realised that this was a unique opportunity to make the material available to a broader audience.

Lessons learned

Without the help by the University of Groningen Press, I would not have published the lecture notes in this form. The barrier to contact and collaborate with a commercial publisher would have been too high.

UG Press and "Uitgeverij kleine Uil" arranged for the publication in a very smooth process, making it a very pleasant experience.

The financial support from the Open Access Textbook Fund oft he University was instrumental. In my opinion, it should certainly be re-installed as an effective tool to facilitate open access publication.

URLs, references and further information

Last modified:13 November 2024 11.33 a.m.