Going from Neurons to Whole Brains
- The Blue Brain Project, proposed, quite some time ago, to
simulate the brain, neuron by neuron.
- The HBP was initially quite related to this.
- If you're neurons are simple (LIF), and you don't need to
do it in real time,
there's no particular reason you can't do it now.
- The problem is that the brain that is simulated doesn't really
do much.
- What is clear is that the current state of the art in computational
neuroscience shows that we don't really know what the brain is doing.
- I think we can make a lot of progress by modelling agents in
real time, which you can do with small brains (billion neurons) in the
HBP.
- Even there, we don't really have learning working very effectively.
- It's not deep learning from AI.
