chase shimmin

I'm a postdoctoral researcher doing physics & machine learning @Yale and CERN/ATLAS. He/him.

[chase shimmin @cshimmin]

@lukasheinrich_

[chase shimmin @cshimmin]

@douglasduhaime 😮 Doug someday you'll have to tell me what this is all about.

[chase shimmin @cshimmin]

@KyleCranmer @nyuniversity I sometimes wonder if I got COVID on the Metro North train home from this. At the time it was still supposed to be "only in China" but I never had such a bad cold in my life!

[Douglas Duhaime @douglasduhaime]

I've been having fun using machine learning to make chiptune audio. I wrote up a little hello world guide in case others want to experiment too :)

https://t.co/xna864VTFW

[Matt Ferraro @mferraro89]

Have you ever wondered why you can't multiply and divide vectors? It turns out you can, they just don't teach it in school!
https://t.co/FvCgrNZcGC

[chase shimmin @cshimmin]

@DanielWhiteson @DanielandJorge "Caught mid extrusion"... did that happen when congress pulled the plug on the whole thing?

[chase shimmin @cshimmin]

@johncarlosbaez @gro_tsen Correct, at low wimp masses, the energy transfer to the target nucleus becomes too small to detect. For higher masses, the local DM energy density is known to be ~0.3 GeV/cm3, so as mass increases, DM particle density decreases, limiting the "luminosity" of the experiment

[chase shimmin @cshimmin]

@gfodor As a professional physicist (not sure if hardcore) I wasn't sure whether I should feel attacked. But then I re-read... can you link to a definition of physist? Google tries to correct me and if I insist on this term I only find results where people have misspelled physicist

[Andrew Gordon Wilson @andrewgwils]

This article also reinforces the common misconception that deep learning is all about flexibility. If that were true, we would only need RBF kernels, which are like infinite width neural networks. In actuality, inductive biases play a large role in the success of deep learning.

[chase shimmin @cshimmin]

@BasicScienceSav "History of philosophy without any gaps" podcast is great! But it is LONG. The first ~40 eps are a nice intro of everything up to Aristotle though.

[chase shimmin @cshimmin]

@johncarlosbaez Hmm I guess not? But it could be related to some kind of topological space.

[chase shimmin @cshimmin]

@johncarlosbaez Out of curiosity, is there a name for the denominator term? It kind of looks like a 2-norm turned inside out.

[Matplotlib @matplotlib]

OMG we're album art!

[Janelle Shane @JanelleCShane]

WHAT HAS OCCURRED CANNOT BE UNDONE

I have trained a neural net on a crowdsourced set of vintage jello-centric recipes

I believe this to possibly be the worst recipe-generating algorithm in existence

[chase shimmin @cshimmin]

@BasicScienceSav Didn't they try this with cryptography and it didn't really work? It just made things a pain for companies, i.e. Java couldn't ship w/ crypto libs so you had to download open source ones separately

[chase shimmin @cshimmin]

@KyleCranmer Vidyo conference

[Shashi Thutupalli @stpalli]

From @skdh’s brilliant “Lost in Math”. All physics curricula should include these lines, which should also be inscribed in all physics departments as a reminder.