DropConnect in LVQ networks for regularization and confidence interval estimation
- DropConnect (the generalization of Dropout) is a very simple regularization technique that was introduced a few years ago and has become extremely popular because of its simplicity and effectiveness. In this thesis, a suitable architecture for applying DropConnect to Learning Vector Quantization networks is proposed along with a reference implementation and experimental results. Inmany classification tasks, the uncertainty of themodel is a vital piece of information for experts. Methods to extract the uncertainty and stability using DropConnect are also proposed and the corresponding experimental results are documented.
Author: | Jensun Richie Dinesh John Ravichandran |
---|---|
Advisor: | Thomas Villmann, Marika Kaden |
Document Type: | Bachelor Thesis |
Language: | English |
Year of Completion: | 2018 |
Granting Institution: | Hochschule Mittweida |
Release Date: | 2019/04/30 |
GND Keyword: | Vektorquantisierung; Regularisierung |
Institutes: | Angewandte Computer‐ und Biowissenschaften |
DDC classes: | 621.3822 digitale Signalverarbeitung, Vektorquantisierung |
Open Access: | Frei zugänglich |
Licence (German): | Urheberrechtlich geschützt |