Pattern recognition in Spiking Neural Nets using Brian2

Hello Brian2 users,

In order to learn Brian2 and build more efficient Spiking Neuron models, I would like to work on implementing simple neuron models used for pattern recognition from research papers. This would benefit both the engineering aspect of coding and modelling and the research aspect of reading the literature.

I would like to propose this paper to implement using Brian2 which looks simple (hopefully) to begin with:
Pattern recognition with Spiking Neural Networks

Brief details from the paper:

Task: Classify an image as either X or 0
Neuron model: The paper uses Izhikevich neurons
Learning rule: STDP is used as the learning rule for training
# of Trials: The authors have simulated for 1000 trials and calculate the accuracy metrics.
Network Architecture: 25 Input Neurons, 5 neurons for hidden layer and 2 output neurons

I would be interested in forming teams to work on this project. My timezone is CST and am looking forward to hearing from you all and learn from this wonderful community.


Hi @touches,
Learning by doing is great, I am interested in the topic, but I have no experience in pattern recognition. I don’t know how much I could be helpful for the team, but I like to follow it.

No worries, the task is simple: to classify the input as either X or 0. Observing the firing pattern for the inputs generated by the Spiking nets one can make classification. I made this post just to see if there are interested folks to form teams an if not I’ll be working on my own and get help from this forum

Awesome, I play around the paper.

1 Like

Sure, I’ll post updates and link to the colab notebook

I want to work on this project. But I haven’t worked on SNN and pattern recognition before. I hope it will give me a good experience to understand how Brian works and what’s SNN. I will start with STDP.

1 Like