It will widen your horizon, they said. And here I was, foolishly thinking I could get away with half-assing statistics during my degree.
As somebody with a degree in bioinformatics, I have never seen something more true in my whole life.
Some more lies from my time in academia:
Biostatistics was the only part of my biomedical engineering PhD course load that I enjoyed
Guess who doesn't have a PhD
My dog
Your dog and I have a lot in common
Almost every Baby Boomer ever.
I thought the punch line was that biostatistics is actual biology, and biology is statistics :)
A someone not in the field (CS/Machine learning) what did you expect these to be?
Whatever course you do in STEM, you don't want to half-ass the first semester of calculus, linear algebra and statistics.
In fact, you probably want to go out of your way to actually learn linear algebra (because I've never seen anybody really learn it on the course, you need to apply it) and statistics (because you want to go deeper).
Linear algebra, absolutely. But I kind of hoped to get through my whole degree (mostly EE) without properly knowing statistics. Hah. First I take an elective Intro AI class, and then BioInf. I guess I hate myself.
Oh you can get through most degrees without properly learning linear algebra or statistics.
But those 3 are the knowledge that will pop here and there on everything you do, and leave you confused, incapable of understand things, and incapable of extending things if you don't know them. Usually, you won't even have to calculate anything, you just have to know them.
Well you need a good handle on probability to understand transistors these days at least.
EE wizards are also the keepers of information theory, which is all just probability, and is pretty much the definition science of our modern era.
Whatever course you do in STEM, you don’t want to half-ass the first semester of calculus, linear algebra and statistics.
shenMask.Jpg
mask label : STEM
face label : Maths
People here saying this shit is useless are fucking wack, I use this shit frequently in my job. Basic affine transforms for grid data is an interview question we ask junior engineers.
My wife did more statistics in her psychology degree than I did in engineering.
Engineering formulas be like
"So there was this guy in 1896 and he did a bunch of trials and he figured out that a+b*x/c² is close enough to the real results, with values for a in range 1-2 and b in range 3-4. We still don't understand why, or how he got there, but it worked ever since."
Scientists want to understand things. Engineers don't care, as long as it works.
"More recent research has produced a more precise emperical relation, -4.2x^3.761+√(sin(2x²/π))±erf(e²ˣ+37), which produces results which are 0.2% more accurate on average."
Because engineering is precise, measurable, and easily reproducible. You should be testing your things in a way that all you need is a simple two-sample Z-test.
Experiments on the humans, on the hand, unfortunately, have been outlawed. So all you get is a bunch of shitty noisy data, and yet you're supposed to somehow make sense of it. Most people with a degree in stats would tell you not to even try, and yet those fucks at phycology departments always do while having had about one undergrad-level class as part of their masters.
TL;DR good psychology programs nowdays train decent statisticians as they should.
It always got me that the maths I was doing in electrical engineering outclassed what my friend was doing for his astrophysics degree. He was probably at the better university too (Debatable for the subjects in question, but both really good).
Did I need that level of maths? No, but it was compulsory for the first 3 years so not much option.
But... but... these are my maths shoes!
Damn, looks like an elaborate course in probabilistic graphical models, sign me up!
I have a degree in regular informatics and I don't really know what that even means. 🤷
A place for majestic STEMLORD peacocking, as well as memes about the realities of working in a lab.
Rules
This is a science community. We use the Dawkins definition of meme.