Kaspersky Neuromorphic Platform architecture

Kaspersky Neuromorphic Platform includes the following components:

• Python framework. It is designed for creating spiking neural networks and converting artificial neural networks into spiking networks. Resulting spiking neural networks are trained with synaptic plasticity methods and can be executed on a central processor, an AltAI-1 neuromorphic processing unit, or a software emulator of the latter. Artificial neural networks converted to spiking networks are trained through backpropagation that considers the limitations of the AltAI-1 neuromorphic processing unit. Resulting neural networks are adapted for execution on the AltAI-1 neuromorphic processing unit, or a software emulator of the latter.
  

  A neural network based on artificial neurons that perform an instantaneous summation of input signals represented as numbers and weighted according to the weights of input connections, and generate an output numerical signal according to a given activation function.

  

  A neural network based on artificial neurons that perform temporal summation of input signals represented by spikes and weighted according to the weights of input connections, and generate an output pulse according to a certain threshold-based activation function.

  The Python framework includes modules whose functionality overlaps with that of the C++ framework and an ANN2SNN package required for converting artificial neural networks into spiking neural networks and executing these on an AltAI ANN2SNN backend.

  

  A backend for training and execution of artificial neural networks (ANN) converted to spiking neural networks running on an AltAI-1 neuromorphic processor.

• C++ framework. It is designed for creating spiking neural networks and training these with synaptic plasticity methods. Resulting neural networks can be executed on a central processor, an AltAI-1 neuromorphic processing unit, or a software emulator of the latter.
• Backend. Starts training and execution of a spiking neural network on various computing devices (for example, neuromorphic processor AltAI-1). The platform provides several independent backends, each letting you train and execute spiking neural networks on an appropriate device type.

The figure below shows a diagram of the interaction between the Kaspersky Neuromorphic Platform components.

Kaspersky Neuromorphic Platform architecture

In this Help section Neural network description format Data exchange mechanism in Kaspersky Neuromorphic Platform Data formats processed by the platform About Kaspersky Neuromorphic Platform backend

Page top