Skip to content
DevMeme
3050 of 7590
Trying to dive into machine learning until mathematics stops me cold
AI ML Post #3365 · source on Telegram

Trying to dive into machine learning until mathematics stops me cold

Description

Two stacked movie-still frames form a meme. In the first frame, a casually dressed man exiting a modern glass-door café is labeled with large white text “Me”; he clutches his chest while admiring a woman walking past who is labeled “Machine Learning.” In the second frame, a different man in a plaid shirt plants his hand firmly on the first man’s chest, blocking his path; this blocker carries the white label “Mathematics.” The background shows polished metal doorframes, reflections in the glass, and urban sidewalk scenery. The joke plays on developers’ eagerness to jump into trendy ML tooling only to be confronted by the prerequisite of solid mathematical foundations - linear algebra, calculus, and probability - before making meaningful progress

Categories

AI ML Mathematics Learning

Common Tags

MachineLearning AI DeveloperHumor LearningCurve RelatableHumor

Comments

30
Anonymous ★ Top Pick Machine Learning looks like the perfect mid-career fling - until mathematics shows up as the bouncer demanding the linear-algebra ID I misplaced around JDK 1.4
  1. Anonymous ★ Top Pick

    Machine Learning looks like the perfect mid-career fling - until mathematics shows up as the bouncer demanding the linear-algebra ID I misplaced around JDK 1.4

  2. Anonymous

    After spending three months debugging a neural network that wouldn't converge, you finally understand why your linear algebra professor kept saying 'this will be useful someday' - turns out, eigenvalues weren't just academic torture, they're why your PCA is crying and your optimizer is having an existential crisis

  3. Anonymous

    Every ML engineer's journey: spending months tuning hyperparameters with frameworks they barely understand, only to realize that Mathematics has been standing there the whole time, patiently waiting to explain why their model actually works - or more often, why it spectacularly doesn't. Turns out you can't just import numpy and call it a day when your gradient descent is converging to the wrong local minimum because you skipped the calculus lecture on convex optimization

  4. Anonymous

    Everyone wants to skip to 'import sklearn; model.fit', but the moment it diverges, the linear algebra bouncer demands ID: eigenvalues, gradients, and a convergence proof

  5. Anonymous

    About to “do ML,” then Mathematics grabs me: “Define the loss, check identifiability and condition numbers, and stop pretending random_state=42 is a methodology.”

  6. Anonymous

    ML: the shiny microservice. Math: the monolith dependency you can never migrate away

  7. Deleted Account 5y

    но математика ведь топ😢

    1. @RiedleroD 5y

      please speak english in this chat or provide a translation alongside your original text

    2. @RiedleroD 5y

      also, I do agree with that statement (maths is top)

    3. @p4vook 5y

      "Слишком математично, я зря родителям говорил, что математика в программировании не нужна и что я справлюсь без универа...." "Too mathematical, I was wrong to tell my parents that mathematics is not needed in programming and that I can do it without a university degree ...."

      1. @RiedleroD 5y

        mate

        1. @p4vook 5y

          bruh google translate got stuck

          1. @viktorrozenko 5y

            Those are chat rules. If you violate them, you'll get kicked. I didn't make 'em

            1. @p4vook 5y

              no problem

            2. @RiedleroD 5y

              and you don't enforce them either - please don't argue with people over it when a mod is here

              1. @viktorrozenko 5y

                I'm not arguing, just explaining, bro

                1. @p4vook 5y

                  Nah, that didn't look like an explanation, but nvm, I got it

                  1. @viktorrozenko 5y

                    Sorry if I sounded harsh. That wasn't my goal

                2. @RiedleroD 5y

                  also not needed - that's what us mods are for. @pavel_the_best knows the rules, he's been active in this chat for quite some time.

                  1. @viktorrozenko 5y

                    Cool, ok

      2. @RiedleroD 5y

        speak English or provide a translation

      3. @p4vook 5y

        that's pinned message in a Russian chat about algorithms

        1. @RiedleroD 5y

          which chat?

  8. Deleted Account 5y

    oh, fuck

    1. @RiedleroD 5y

      no worries, just remember this rule :)

  9. Deleted Account 5y

    dude, I understand u

  10. @fffff80240d322f9 5y

    I think it's mainly statistics right? Or are there other branches of math that ml uses?

    1. @karim_mahyari 5y

      If you're going for regression or classification models, you might want to review your linear algebra and calculus notes Much of the math there is based on matrices and partial derivatives

    2. @yarik15886 5y

      Also calculus and linear algebra

  11. @kral_na_severu 5y

    learn js, make money without math

More like this

More like this in AI ML →
AI ML MachineLearning AI
Stuffing Kimi K3 2.8T Into a 3090

Stuffing Kimi K3 2.8T Into a 3090

AI ML
One Prompt Ends an Eighty-Seven-Year Conjecture

One Prompt Ends an Eighty-Seven-Year Conjecture

Mathematics
The Alpha Model Skips Retrieval Entirely

The Alpha Model Skips Retrieval Entirely

AI ML
Use J and K for navigation