Léo Varnet @leovarnet@qoto.org

DAGs, Golems, and Owls: Statistical Rethinking 2023 Lecture 1 (of 20). No hard work in this introductory lecture, just a conceptual outline and some dank memes. Lecture 2 later this week introduces Bayesian inference. https://www.youtube.com/watch?v=FdnMWdICdRs&list=PLDcUM9US4XdPz-KxHM4XHt7uUVGWWVSus&index=1

Another very interesting piece in this museum is the Mini-componium, a replica of the original Componium built by Diederich Nikolaus Winkel in 1821. This remarkable mechanical music instrument can vary the music it plays using a randomizing mechanism! In fact, the Componium creates endless variations on a theme by selecting randomly one of the 8 possible melodies every two measures.
In a sense, this is a very early example of
#ComputerMusic #AlgorithmicComposition
(https://en.m.wikipedia.org/wiki/Componium)
(3/3)
#musicbox #music

My periodic reminder that my book on the discrete Fourier transforms (incl. FFT, finite fields valued FT, representation theory, and more) is available for free. https://mathematical-tours.github.io/daft/

These music boxes were probably built at the end of the 19th century. An extra feature is the panorama of the horse race in the cabinet. By inserting a coin, the music is played and one can gamble on the winning horse. The machines were designed in such a way that the winning horse could not be predicted with any certainty.

I don't know how it works, but this is maybe one of the earliest random number generator! (2/2)

If you use frequentist statistics, you have to accept that 'no isolated experiment, however significant in itself, can suffice for the experimental demonstration of any natural phenomenon; for the "one chance in a million" will undoubtly occur, with no more or less than it's appropriate frequency, however surprised we may be that it should occur to us.' (Fisher). Replication is a core requirement of all frequentist statistics.

New paper provides a history of “voodoo science,” which discusses the controversy surrounding Vul et al.’s (2009) controversial article “Puzzlingly High Correlations in FMRI Studies of Emotion, Personality, and Social Cognition.”

Five quotes follow: 🧵👉

🔓 https://doi.org/10.3390/socsci12010015

#MetaScience
#Neuroscience
#Neuroimaging
#MetaResearch
#PsychMethods
#ReplicationCrisis
#PhilosophyOfScience
#PhilSci
#Fmri
#VoodooCorrelations
#UseNovelty
#MultipleTesting

There’s an absolutely astonishing story brewing in the medieval manuscripts world of a Potemkin Village scholarly institute built out of plagiarism and stock photos. I can scarcely do it justice except to link to Peter Kidd’s thorough deconstruction of #receptio #receptiogate

Part 1: https://mssprovenance.blogspot.com/2022/12/the-receptio-rossi-affair-part-i-staff.html

Part 2: https://mssprovenance.blogspot.com/2022/12/the-receptio-rossi-affair-part-ii.html

Part 3: https://mssprovenance.blogspot.com/2022/12/the-receptio-rossi-affair-iii-my.html

Part 4: https://mssprovenance.blogspot.com/2022/12/the-receptio-rossi-affair-iv-my.html

Part 5: https://mssprovenance.blogspot.com/2022/12/the-receptio-rossi-affair-part-v.html

Ce sera bientôt les 10 ans de la disparition du regretté Aaron Swartz, le 11 janvier 2013 à l'âge de 26 ans.

Trop peu connu des élèves, étudiants et professeurs, je me suis permis d'y recopier ci-dessous le clip réalisée par France Culture en 2020 qui constitue une excellente entrée pour découvrir sa vie, son œuvre et ses luttes qui demeurent ô combien d'actualité.

Le 11 janvier prochain, parlons d'Aaron Swartz tout autour de nous !

https://www.radiofrance.fr/franceculture/aaron-swartz-hacktiviste-des-savoirs-4602698

#AaronSwartz #AaronSwartzDay

Au passage, ce petit exemple souligne le fait que la séparation entre les langues accentuelles (dont le français) et les langues à tons est plus subtile qu'on pourrait le penser de prime abord : en français aussi la mélodie sur laquelle on prononce une phrase peut en changer radicalement le sens. #psycholinguistique #linguistique

A thread on ignorance in science.

It’s often claimed that science is about accumulating knowledge. But knowledge accumulation is only one side of the scientific coin. This thread contains 10 quotes from scientists, philosophers of science, sociologists, feminists, and statisticians who’ve highlighted the importance of *ignorance* in science, both as a precursor and a product. Enjoy!

#Science
#PhilosophyOfScience
#PhilSci
#SociologyofScience
#Statistics
#Feminism
#Knowledge
#Ignorance

1/11 🧵👉

"It seems astounding to me that the idea of correspondence between brain activity and ambient features has ever been taken seriously. To see a neuron as being responsible for a percept ... is straight vitalism or animism: It attributes to one component of a system all of the properties of a description..."

"...If a car undergoes a structural change because of bumping into a tree (the fender is twisted and the front tire scratched), we do not say that it remembers the accident by storing a memory, nor do we say that it learned from the event by changing its behavior via a representation of trees. These descriptions seem ludicrous because the car is so obviously a man-made artifact. Yet it seems that the same sort of mechanistic and operational description can be applied to highly plastic autonomous systems like the nervous system. Memory requires no record or storage, for it stands only for a history of structural coupling; learning requires no representation, for it stands for structural plasticity. Whatever the observer wants to see or needs to use in symbolic descriptions, such as storage, and representation as mapping, are not operational."

Francisco J. Varela
Principles of Biological Autonomy (1979)

#FranciscoJVarela #Memory #SymbolicRepresentation #biology #cognition #Cybernetics #philosophy @philosophy

P. 259

Popper on 'best theories'. As opposed to what people often think, Popper is clear theories often have 'ad hoc' assumptions introduced to 'fix' a theory when original predictions do not pan out. He just prefers ad hoc assumptions to be testable and as few as possible. If your preregistered analysis seems like a bad idea afterwards, you can ad hoc your analysis. It is better if you don't have to, and you should not do it to escape falsification. But it can be fine (and future tests will tell us).