Thread by @jslbutler

@neuromatch starts tomorrow with a Python refresher.
I’m going to give a musical pairing for each day.
For W0D1 my suggestion is Holst: The Planets.
In particular, Jupyter: the Bringer of Joy.
Best of luck to all the students and TAs I hope it brings joy
open.spotify.com/album/4v0Xyz0L…

@neuromatch My W0D3 @neuromatch musical pairing for @EllaBatty Linear Algebra is the excellent Tron Legacy Soundtrack.
The connection is in Tron they use vectors, matrices and rotations to visualise a digital world.
Similar visualisation are useful for motor control😀
open.spotify.com/album/3AMXFnwH…

My W0D4 @neuromatch musical pairing for
@arvin_neuro @EllaBatty and me tutorials on Calculus is My Favourite Things as they are some of my favourite things.
For Tutorial 1 on Differentiation and Integration I recommend the classic Julie Andrews version /1
open.spotify.com/track/1iCHl2R7…

For W0D4 Tutorial 2 on Differential Equations which was written with help from @rebecca_mbrady
I recommend the John Coltrane version as you can hear the connection to original (differentiation) but it is self referential and different themes emerge /2
open.spotify.com/track/3ZikLQCn…

For W0D4 Tutorial 3 on Numerical Methods which was written with help from Harvey McCone
I recommend the Jon Batiste version as it is stripped down and simplified but still shows the magnificence and beauty of the song (differential equations) /3
open.spotify.com/track/2DCItemp…

Thanks to everyone in @neuromatch for all the help with getting W0D4 tutorials ready as writing them has been one of my favourite things of the past year. /end

My W0D5 @neuromatch musical pairing for @UlrikBeierholm tutorials on Probability and Statistics is the musical statistical summary songs by the now award winning statistician/musician @rafamoral
My favourite is the ROCk curve

@UlrikBeierholm An alternative pairing could be that Statistics is "Public Enemy" of some researchers and you have to "Fight the Power" (analysis) :)
open.spotify.com/track/1yo16b3u…

My W1D1 @neuromatch musical pairing for Model Types by @GunnarBlohm @KordingLab @bgalbraith is
"The Model" by Kraftwerk, 1970s electronic music at its best.
With the addition of four songs for three thematic questions below👇1/2
open.spotify.com/album/3eyz60xE…

"What a Wonderful World" by Louis Armstrong
"How Bizarre" by OMC
and
"Why" by Annie Lenox
or
"Why" by Bronski Beat
Because both songs are great and it is always worth asking why twice.
2/2
open.spotify.com/track/7vOLJwP5…

My W1D2 @neuromatch musical pairing for Model Practices by @BMtHart @meganakpeters @schrater_paul @GunnarBlohm is "1234" by Feist.
For the 4 first steps of modeling 1/2
open.spotify.com/track/3GMlL33B…

And "Think" by Aretha Franklin
Because you need to Think about how to frame your question 2/2
open.spotify.com/track/4yQw7FR9…

My W1D3 @neuromatch musical pairing for Model Fitting by Pierre-Étienne Fiquet, @anqiwu_angela @AlexHyafil @jdrugowitsch is "Largo al factotum della città" from The Barber of Seville.
Because like Figaro a Linear Regression model can fit anything (linear)
open.spotify.com/track/41PYCIme…

Full disclosure I've used this analogy before and I still stand by it

https://twitter.com/jslbutler/status/1390708019210641418?s=20

My W1D4 @neuromatch musical pairing for Generalized Linear Models by Cristina Savin, Pierre-Etienne Fiquet @arisbenjamin @jakhmack
is "I am the very model of a modern major general",
because GLMs "understand equations, both the simple and quadratical"

My W1D5 @neuromatch musical pairing for Dimensionality Reduction by @YuLikeNeuro Alex Cayco Gajic, @johndmurray are "One" by Aimee Mann and "Soy Yo" by Bomba Estero.
One because you want to reduce those dimensions down to one (well maybe two or more) 1/2
open.spotify.com/track/5e7OZ8Py…

And "Soy Yo" because you have to know yourself to know your Eigenvectors and Eigenvalues. 2/2
#neuromatch
open.spotify.com/track/4Egb5xP6…

My @neuromatch musical pairing for Week 1 professional development is "9 to 5" by Dolly Parton.
open.spotify.com/track/4w3tQBXh…

My @neuromatch musical paring for Deep Learning by Aude Olivia, Jorge A. Menendez, @computingnature and Yalda Mohsenzadeh are:
"Expensive Shit" by Fela Kuti
"Anything Goes" performed by Sutton Foster
Because Deep Learning can be computational Expensive 1/3
open.spotify.com/track/5FvggTB5…

And a good deal of people think the with Deep Learning "Anything Goes" but picking the right tool is very important
2/3

https://twitter.com/BMtHart/status/1414367853315194881?s=20

And if you need a pick me up on Monday morning of week 2, this video of Sutton Foster performing "Anything Goes" just puts a smile on my face 3/3

My W2D2 @neuromatch musical pairing for Linear Dynamical Systems by Eric Shea-Brown, @bingbrunton and Ellie Stradquist are
"Should I Stay or Should I Go" by the Clash &
"Stop and Go" by the Regrettes
Because "this indecision's bugging me"... 1/2
open.spotify.com/track/39shmbIH…

.. so I model it with dynamical systems with stochasticity like the drift diffusion model which can be extended to model "Stop and Go". 2/2
open.spotify.com/track/1H1wMFBD…

My @neuromatch musical pairing for Biological Neuron Models (W2D3) by Upi Bhalla, @qinglong_gu, Songtin Li, @johndmurray @NeuroNaud @arvin_neuro @YiotaPoirazi is
"Even Better Than The Real Thing" by U2
because some models can be better than the real thing

My @neuromatch musical pairing for Dynamic Networks (W2D4) by Nicolas Brunel, @qinglong_gu Songtin Li, @arvin_neuro @johndmurray @GjorJulijana & Ken Miller is
"Don't Give Up" Peter Gabriel & Kate Bush
because the Wilson-Cowan model is a beautiful duet 1/2

between inhibition and excitation, like the song.
And we are just after the mid-point of the @neuromatch Academy energy levels can be dipping but do not give up and do not worry if some of the materials seem hard they can take some time to sink in 2/2

I'm delighted to be doing a couple of @neuromatch Q and As on Dynamical Systems today, the panels are excellent with @bingbrunton @GjorJulijana @arvin_neuro et al.
In comparison I feel like Ringo in Beetles but he did sing yellow submarine so that is good!
open.spotify.com/track/1tdltVUB…

My W2D5 @neuromatch project day musical pairing is the album “Dirty Computer” by Janelle Monae.
Because if your code does not work it is your dirty computer’s fault, it is definitely not a line in your code, definitely not, well maybe it is your code.
open.spotify.com/album/2PjlaxlM…

My @neuromatch musical pairing for Bayesian Decisions (W3D1) by @schrater_paul @NeuroChooser @xaqlab @EllaBatty Saeed Salehi is:
"Gone Fishin" by Louis Armstrong & Bing Crosby.
It will make sense when you do the tutorial.
open.spotify.com/track/4CtfrD23…

The @neuromatch musical pairing for Hidden Dynamics (W3D2) by @seanescola, Yicheng Fei, @xaqlab, @JesseLivezey & Itzel Olivos Castillo is:
"Hide and Seek" by Imogen Heap
As sometimes we don't know "What the hell is going on?" because it is hidden. 1/2
open.spotify.com/track/7mMlbJlX…

And for the musical pairing for @neuromatch (W3D2) Hidden Dynamics Bonus materials by Caroline Haimerl @bgalbraith & Maryam M. Shanechi here are the hidden states of the song "Hide and Seek".
Great musical pairing suggestion by @xaqlab and @schrater_paul
tiktok.com/@sampopemusic/…

@neuromatch musical pairing for Optimal Control (W3D3) by Maurice Smith, Zhengwei Wu, @shreyaneuroctrl @xaqlab & Reza Shadmehr are
"Learning to Fly" by the Foo Fighters &
"My Shot" by Lin Manuel Miranda

@neuromatch musical pairing for Reinforcement Learning (W3D4) by Doina Precup, @marcelomattar, @NeuroChooser @prokraustinator @bgalbraith & @behrenstimb are "Oompa-Loompa" from Willie Wonka &
"True Faith" by New Order (suggested @bradpwyble)
because 1/2
open.spotify.com/track/1yRfCCqy…

to "seamlessly" mash these the lyrics together:
Oompa loompa doompety da
If your models not greedy, you will go far
your model will converge in happiness too
while you think that the day would never come
Spend my life in the shade of the morning sun 2/2

@bradpwyble did you see what I did with the lyrics almost as genius as my Jupiter, Jupyter pun early on, I am definitely not loosing steam.

@threadreaderapp unroll