Sergio Pecirno

Sergio Pecirno

Computational behavioral neuroscience PhD student in the Keinath Lab at the University of Illinois at Chicago Brain and Cognitive Sciences, Psychology.

HTML versions of my publications can be found here, and an up-to-date list of my publications can be found on Google Scholar. Other stuff I'm working on can be found in my blog. For the Keinath Lab members, calcium imaging code can be found in the calcium imaging code repository on GitHub.

MNIST: I trained a 3 layer MLP with a 2 neuron bottleneck layer (l2) (W1: 784x128 W2: 128x2 W3: 2x10).I trained the model for 100 epochs with a sigmoid activation function around W1, W2 + biases. I then ran 10,000 test examples and grabbed the activations from the neurons in the second layer. I wanted to know what neurons in the first layer of the MLP pushed the geometry of the bottleneck activations in particular directions. [l1 neuron inspector]