Author: Keyao Li
Supervisors: Xavier Otazu
Presentation time: 9:15
Virtual Room: 4.2 | Live presentation URL
In this thesis we try to solve a simple classification task using a biologically plausible computational mechanism. We use a spiking neural network (using NEST simulator software) whose learning process is the biological mechanism called Spike Time Dependent Plasticity (STDP). This mechanism changes the connections between nodes of the network depending on the temporal synchronization of spikes. Input data is encoded, similarly to biological cells, using Gaussian receptive fields. We applied this architecture to MNIST dataset, obtaining high accuracy classification results.