Take home lessons

Artificial neural network: Interconnected group of nodes, inspired by a mathematically abstracted model of a biological neural network.

Deep neural network: An artificial neural network wherein groups of neurons are organized in interconnected layers.

Convolutional neural network: A deep neural network that incorporates convolutional and pooling layers. Mostly used for training over matrices (or frames) for visual processing.

Recurrent neural network: A deep neural network that features feedback (or recursive) connections between neurons. Mostly used for training over temporal sequences.

Biological neural network: A neural network comprises biological neurons connected via electrical or chemical synapses.

Spiking neural network: In contrast to ANNs, which communicate differentiable values, SNNs propagate discrete spikes. A spiking neuron produces spikes in a response driven by spikes arriving from other neurons. Incoming spikes are filtered, weighted, summed, and evaluated according to some neuronal dynamics which dictates the threshold for the initiation of output spikes.

Spiking neurons can be organized in layers to construct deep neural networks, wherein each neuron in one layer is connected to each neuron in the successive layer. Each connection has an adjustable weight.