Human opinion is entertaining, data is trustworthy.Bryan Johnson

Without data you’re just another person with an opinion.William Edwards Deming

Without insight into data, people make decisions based on instinct, speculation, or prevalent theory. People are at risk of acting on biases or false assumptions.


https://www.youtube.com/playlist?list=PLSQl0a2vh4HC9lvrBhVt4UUkhzpp3N5_x

https://taylor.town/compress-code

https://seeing-theory.brown.edu/

Cause and Effect

Don't fool yourself.

Not all scientific studies are created equal

$$ P(A \mid B)= \frac{P(A \cap B)}{P(B)} $$

$$ P(A \cap B) = P(B \mid A)P(A) = \frac{P(A \cap B)}{\cancel{P(A)}}\cancel{P(A)} $$

Machine Learning, Neural Networks, Deep Learning

say we want to classify into $Ys$ given $Xs$

maybe $Y = 1 = \text{cat}$ and $Y=0 = \text{not cat}$

$$ P(A, B) = P(A \cap B) = P(A)P(B \mid A) = P(B)P(A \mid B) $$

$$ P(Y| X) = \frac{P(Y)P(X|Y)}{P(X)} $$

for classification,

$$ \argmax_\theta(\theta \mid X) = \argmax_\theta \frac{P(X \mid \theta)P(\theta)}{\cancel{P(X)}} = \argmax_\theta P(X\mid\theta)P(\theta) $$

we can just cut $P(X)$ out because it’s not dependent on $\theta$, it’s just a constant that stretches the data.

$P(\text{probability of the \color{red}{data}\color{default}{, given params)}}$

$L(\text{likelihood of the \color{green}{params}\color{default}{, given data)}}$

https://www.youtube.com/playlist?list=PLblh5JKOoLUK0FLuzwntyYI10UQFUhsY9

https://www.youtube.com/playlist?list=PLUl4u3cNGP60hI9ATjSFgLZpbNJ7myAg6

https://youtu.be/h8EYEJ32oQ8

https://youtu.be/uzkc-qNVoOk

https://youtube.com/watch?v=XJnIdRXUi7A

https://www.youtube.com/playlist?list=PLZHQObOWTQDOjmo3Y6ADm0ScWAlEXf-fp

https://www.youtube.com/watch?v=zeJD6dqJ5lo&t=66s

https://www.youtube.com/watch?v=cy8r7WSuT1I

https://youtu.be/6YzrVUVO9M0?si=cNnC0p1ifGW687Sd

https://youtu.be/SkidyDQuupA