CloudBrain is a tool for simulating neural networks. It is based on modern cloud infrastructure and on the Spiking Neural Network model. Typical neuromorphic design goals, such as small form factor and low power consumption, are traded for:
- no constraints on the model elements
- access to all events and parameters during and after the simulation
- online reconfiguration of the network
- real-time simulation.
CloudBrain is open source and can run on commodity hardware or in the cloud, thus providing a platform with a different set of features supporting research into, for example, neuron models, structural plasticity and three-factor learning.
CloudBrain: Online neural computation in the cloud
Leon Bonde Larsen, Rasmus Karnøe Stagsted, Beck Strohmer, Anders Lyhne Christensen
Integrating Non-Spiking Interneurons in Spiking Neural Networks
Beck Strohmer, Rasmus Karnøe Stagsted, Poramate Manoonpong, Leon Bonde Larsen