Why he quit a $500k+ machine learning job at Meta (Facebook): a candid review of his experience, mistakes, and ML best practices - Damien Benveniste - the data scientist show049
The Data Scientist Show - Daliana Liu
00:00
Do You Have a Base Line for Machine Learning?
I think one thing is always important when we develop much learning model is to know your base line. So comparent to the base line, ah, how does your model work? Right? Maybe i can just improve ten % compared to the base lines. Or maybe you develop a model that has 95 % accuracy, but the human labelling ot whatever base line is already pretty high, is 90 nive%. At that time, maybe you need to think about, oh, is this really worth it to invest in muchin learning? I think knowing the base line and setting a goal comparing to the baseline is a way for us to, you know, set a milestone, create a mean
Transcript
Play full episode
