-
Why second-order methods can be futile in non-convex problems
I’ve been working on fitting a convolutional model of neurons in primary and intermediate visual cortex. A non-convex optimization problem must be solved to estimate the parameters of the model. It has a form similar to: There are some shared weights to further complicate things, but the most salient features is that it’s a 3-layer
-
The secret ingredient in stochastic gradient descent
I had dinner with Geoffrey Hinton and Yoshua Bengio a few weeks back, and I left full of ideas – and wine, also. Now I’m fitting a massive model for early and intermediate visual areas which involves major spiffiness and about 100 hours of data (!). Stochastic gradient descent is one optimization algorithm which can
-
Try Zotero Reader for tablet
Update: Zotero Reader Web is now freely available. Zotero is a citation manager that I adore. Previously, I lamented that it’s troublesome to sync a Zotero library to a tablet. I’ve just finished a simple app to give you an easy way to view your Zotero library on the go. Kind of like Pocket (formerly
-
Category representation in the brain
Jack Gallant has been focusing a lot of recent lab efforts in the analysis of large-scale fMRI recordings. They published an interesting paper in Neuron last month about the representation of categories in cortex.
-
Killer among us – Dr. Fabrikant
In the wake of all the talk of gun control re: Sandy Hook, I looked into the Concordia University massacre. It was big news in Quebec in 1992, and I remember hearing about it as a kid. I think many of you will be interested in the case as the shooter was a university professor
-
Temporal impulse response functions of LFPs
I’m about to submit a paper on LFP receptive fields in V4 and in the interest of flow had to scrap a couple of figures. We measured temporal impulse response functions of LFPs using a standard sparse flashed bar paradigm. We used two Utah arrays to measure the signals (96 electrodes arranged in a 10×10
xcorr: AI & neuro 