We created a fork of DEAP ( http://deap.readthedocs.org/) the python evolutionary algorithm framework with an implementation of the IBEA selector ( http://www.tik.ee.ethz.ch/sop/pisa/selectors/ibea/ibea_documentation.txt ).
This selector is particularly useful in the case of multi-objective optimizations.
Would it make sense to combine this with the “Simplification of detailed models” hackathon activity? It sounds as though this would be useful in fitting the AdExp model to detailed NEURON cell types.
Hi,
Yes, it would definitely be useful. Let’s see if some other people propose other problems they want to optimize (like synaptic parameters etc.). Fitting simplified models would definitely be a good use case for DEAP optimisation.