My old buddy Simon Barthelmé as well as Nicolas Chopin and Adam Johansen are organizing a 3-day workshop in Warwick, UK in September on statistical challenges in neuroscience. Some of the invited speakers include J.D. Victor, Robert Kass and Aapo Hyvärinen. Great chance to discuss and present your latest methods-heavy research. Here’s the pitch:
Workshop announcement: NeuroStats 2014 (3-5 September, 2014, Warwick University, UK)
NeuroStats 2014 (Statistical Challenges in Neurosciences, 3-5 September, 2014 , Warwick University, UK) aims to bring together neuroscientists (interested in methodological issues) and statisticians (interested in new applications) with the goal to (a) introduce the neuroscience community to recent developments within statistics and (b) raise awareness within the statistics community of some of the challenges that arise in the analysis of neuroscientific data.
More information may be found at:
http://www2.warwick.ac.uk/fac/sci/statistics/crism/workshops/neuroscience/
Due to practical constraints, the workshop is limited to around 80 participants, so early registration is recommended.
I’ll present my latest research on applying deep neural networks to the study of hierarchical visual processing. Here’s three slides to titillate you:
One layer of the model:
The full network architecture:
And estimated receptive fields in V2:
xcorr: AI & neuro 
2 responses to “Workshop on statistical challenges in neuroscience”
[…] internal nonlinearities in both the excitatory and inhibitory components. So you still need the shared deep neural net magic model™, but at least now you can demonstrate that one of the prominent things it finds – […]
[…] of 0.38 – which brings us much closer to the score in V1. Furthermore, using a fancier model, which I reported briefly on earlier and will publish soon, I got to an average validated r ~ 0.52. In other words, there is a lot of […]