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
-
25 April 2025
Leading microbiologist Arnold Driessen honoured
On 25 April 2025, Arnold Driessen (Horst, the Netherlands, 1958) received a Royal Decoration. Driessen is Professor of Molecular Microbiology and chair of the Molecular Microbiology research department of the Faculty of Science and Engineering at the...
-
24 April 2025
Highlighted papers April 2025
The antimalarial drug mefloquine could help treat genetic diseases such as cystic fibrosis, Duchenne muscular dystrophy, as well as some cancers.
-
22 April 2025
Microplastics and their effects on the human body
Professor of Respiratory Immunology Barbro Melgert has discovered how microplastics affect the lungs and can explain how to reduce our exposure.