-
Notifications
You must be signed in to change notification settings - Fork 0
Home
mundya edited this page Sep 19, 2014
·
1 revision
Paper to discuss the algorithm for Nengo/SpiNNaker, needs to illustrate:
- Efficient use of the SpiNNaker platform (processing/memory usage, network utilisation)
- Description of simulation algorithm (equivalence to spike transmission)
- Description of learning algorithm
- Profiling: speed, memory usage, loading time, (brief) network utilisation
- Applicability to other neural network types? Can good approximations of sparse matrices be made?
Should be an interesting application of distributed computing. Aim for IEEE Transactions on Parallel and Distributed Computing applications and software sections?