Progress

Questions
#1

What percentage of the mouse retina have we mapped so far?


I am asking that because I believe skilled “sorters” can help us map more efficiently, by quickly identifying cubes that would most benefit from individual human attention. 

I have posted some before and after pics on Facebook. When I looked at each cube for the first time, I predicted whether it was a Prom (promising), Pleco or slug (minimal patching expected), or Snowball('s chance in Haleyville of yielding significant human discoveries). I don’t know all the technical hiccups, but I’d like to see us use this predictive ability to move forward a little bit faster maybe.
#2

If I have not misunderstood the purpose of our task, the question is: how the AI program that traces neurons has learned from our work?

Ciao
Mario
#3

@beckeemo So far we’ve only mapped a small percentage of the neurons in this dataset.  But we don’t have to map them all to get started with science. We are about to launch a major push to map the inputs to the J type of ganglion cell.  The goal is to figure out how the J cell detects motion.  We’ll be talking more about this over the next few weeks.

@mariomar Actually it’s not just about training the AI.  We are also doing neuroscience together.