Statistical physics of learning vector quantization
PhD ceremony: Mr. A.W. Witoelar, 11.00 uur, Academiegebouw, Broerstraat 5, Groningen
Thesis: Statistical physics of learning vector quantization
Promotor(s): prof. M. Biehl, prof. N. Petkov
Faculty: Mathematics and Natural Sciences
The field of machine learning concerns the design of algorithms to learn and recognize complex patterns from data. Learning Vector Quantization (LVQ) constitutes an important family of such algorithms using prototypes which serve as typical examples. Despite its wide range of applications, the theoretical understanding of LVQ in general remains very limited. In this thesis a theoretical framework is constructed using concepts from statistical physics which allows for an exact analysis of the typical learning behaviors of the system. Studies in this thesis compare the characteristics of various LVQ algorithms to demonstrate the robustness of Neural Gas (NG) schemes and the specific advantages of LVQ 2.1 and Robust Soft LVQ algorithms. Furthermore surprising non-trivial behaviors are revealed including learning plateaus and phase transitions in the training process. The results provide insights to general prototype-based learning prescriptions.
Last modified: | 13 March 2020 01.13 a.m. |
More news
-
13 May 2024
‘The colourful cells of petals never get boring!’
Most people will enjoy colours in nature. However, the interest of evolutionary biologist Casper van der Kooi goes much further: he studies how flowers, birds, butterflies, and beetles get their colours. He also studies how these colours are used...
-
13 May 2024
Trapping molecules
In his laboratory, physicist Steven Hoekstra is building an experimental set-up made of two parts: one that produces barium fluoride molecules, and a second part that traps the molecules and brings them to an almost complete standstill so they can...
-
07 May 2024
Lecture with soon to be Honorary Doctor Gerrit Hiemstra on May 24
In celebration of his honorary doctorate, FSE has invited Hiemstra to give a lecture entitled ‘Science, let's talk about it’ on the morning of 24 May