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
-
14 July 2025
ERC Proof of Concept grant for Kottapalli and Covi
Professors Ajay Kottapalli and Erika Covi have received Proof of Concept grants from the European Research Council (ERC).
-
10 July 2025
Dutch Research Agenda funding for nanomedicine research
Prof Dr Anna Salvati, Dr Christoffer Åberg and Prof Dr Siewert-Jan Marrink have been granted a National Science Agenda (NWA) funding to further develop life-saving drugs based on nanotechnology with the NanoMedNL consortium.
-
10 July 2025
Dutch research Agenda funding for circular bio-based materials
Prof Anastasiia Krushynska has been granted a EUR 600,000 National Science Agenda (NWA) funding to help develop innovative technologies for converting low-grade organic waste into durable, recyclable materials.