
I propose we replace UTC with Taylor Time (TT). TT+0 is obviously the same as Eastern Time, as that's where Midnights released at midnight.
Seems like a much more relevant way to determine the canonical time zone than centering it on Greenwich
2
Replying to @DynamicWebPaige
I almost wish reservoir computing had won out over backprop just so that I could include this image of "The Liquid Brain" from Pattern Recognition in a Bucket at the start of all my deep learning talks
2
Stats brain: the std of an exponential distribution is the same as its mean
Viz brain: cool, lets take a look at the error bars on this scatter plot
Viz brain: ...
Stats brain: like I said, the std of an exponential distribution is the same as its mean
5
As it happens, you can use Limma to remove predictive linear signal instead of batch effects. When we did so, we found that logistic regression's performance becomes random (as expected), but the neural nets are still able to learn things. 7/
You're probably thinking, "Well you just said the model behavior could be unstable, so what?" But these results held across datasets and problem settings. 5/
1
4
Ultimately I decided to narrow the project down to a make-or-break experiment: given a relatively easy problem on clean data, do neural networks do any better than logistic regression?
As it turns out, no.
4/
A few years ago I was an optimistic young PhD student who set out to develop new ways to apply transfer learning to gene expression data. Given the breakthroughs in semi-supervised learning and the millions of unlabeled expression samples, it seemed like a slam-dunk project. 2/
My favorite part about scientific peer review is that in order to do your duty as an unpaid consultant, you must first give the publisher your personal information and allow them sell it
2
Gene expression data is so messy. It's like if everyone's camera used different wavelengths of light
5
Me: "Okay CLIP Guided Diffusion, show me your interpretation of the Pixar movie UP."
Diffusion: "Did you say photorealistic octopus?"
1
1












