Klaverstijn, Jory (2021) Designing a Hopfield Neural Network Using Memristors. Bachelor's Thesis, Artificial Intelligence.
|
Text
Bachelor_Project_Jory_Final.pdf Download (1MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (122kB) |
Abstract
In this paper it is explained how Nb-doped SrTiO$_3$ memristors could be integrated into classical Hopfield neural networks (HNN), and how a working simulated model could be designed. The performance of this model was tested and compared to the performance of a linearised version, non-memristor version and a modern continuous HNN. The performances of the models are measured on the MNIST data-set by classifying the converged patterns with a secondary feed forward network. From the experiment, it can be concluded that a memristor HNN model can have an adequate accuracy. The accuracy of the (linearised) memristor HNN ranged from 63.6\% to 89.4\%, dependent on the amount of noise added to the images. It is shown that the linearised version of the memristor based HNN performs slightly better and that the memristor based HNNs perform worse than the continuous HNN model for higher magnitudes of noise.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Tiotto, T.F. and Borst, J.P. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 07 Jul 2021 10:33 |
| Last Modified: | 07 Jul 2021 10:33 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/25007 |
Actions (login required)
 |
View Item |