Boekestjin, Joppe Wik (2021) Using memristors in a functional spiking-neuron model of activity-silent working memory. Master's Thesis / Essay, Artificial Intelligence.
|
Text
Master_Thesis_Proposal (10).pdf Download (6MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming.pdf Restricted to Registered users only Download (117kB) |
Abstract
In this paper, a spiking-neuron model of human working memory by Pals et al. (2020) is adapted to use Nb-doped SrTiO3 memristors in the underlying architecture. Memristors are promising devices in neuromorphic computing due to their ability to simulate synapses of artificial neuron networks in an efficient manner. The spiking-neuron model introduced by Pals et al. (2020) learns by means of short-term synaptic plasticity (STSP). In this mechanism, neurons are adapted to incorporate a calcium and resources property. Here a novel learning rule, mSTSP, is introduced, where the calcium property is effectively programmed on memristors. The model performs a delayed-response task. Investigating the neural activity and performance of the model with the STSP or mSTSP learning rules shows remarkable similarities, meaning that memristors can be successfully used in a spiking-neuron model that has shown functional human behaviour. Shortcomings of the Nb-doped SrTiO3 memristor prevent programming the entire spiking-neuron model on memristors. A promising new memristive device, the diffusive memristor, might prove to be ideal for realising the entire model on hardware directly.
| Item Type: | Thesis (Master's Thesis / Essay) |
|---|---|
| Supervisor name: | Borst, J.P. and Tiotto, T.F. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Master's Thesis / Essay |
| Language: | English |
| Date Deposited: | 02 Jul 2021 11:20 |
| Last Modified: | 02 Jul 2021 11:20 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/24899 |
Actions (login required)
 |
View Item |