>Research interests: genetic algorithms ITT red flags that

>String theory
>Magnetic monopole

File: 2de7bceb79353cbe69dcc34a51488fe9.jpg (24KB, 300x400px) Image search: [Google]

24KB, 300x400px

>machine learning

>>8497336
Good tier
>legit rigorous research into ML done at stats and some CS departments

Okay tier
>important applications of ML (e.g. autonomous driving, cancer detection)

kys tier
>unfounded attempts to apply ML to make research sound more important and quantitative (e.g. astronomy, social sciences, ecology)

>metabolomics

>>8497308
>posts on sci

Anyone who relies on predictions or algorithms and doesn't work in an actual lab to confirm them

>Research interests: cancer, Alzheimers, diabetes

>>8497308
meme theory

>>8497355
ML for social science and ecology are legitimate applications though. The hard part becomes making sure you're accounting for as many variables as possible while not fucking up with multicollinearity between them, or subsets of them.

>>8498674
>"we applied a machine learning model to our data"
>just did a logistic regression

>>8498680
A logistic regression is technically a machine learning model, though. The model is the equation, and its machine learning when you're using a machine to process large sets of data to form that model.

So I guess its kind of stupid to say that if you're using something with only several data points and independent variables to consider, but ecology and social sciences typically offer huge data sets and many variables to account for.

>>8497333
>String theory

this + graphene

>>8498692
it is technically machine learning.

but people just throw "machine learning" into the abstract of their social sciences paper to seem more legit.

File: 587.jpg (39KB, 600x600px) Image search: [Google]

39KB, 600x600px

>>8497355
>tfw your boss attempts to "apply" machine learning to an astronomy problem just to make it sound flashier