Is machine learning just statistics? | Charles Isbell and Michael Littman and Lex Fridman

Video quality	The size	Download

Information Is machine learning just statistics? | Charles Isbell and Michael Littman and Lex Fridman

Title	:	Is machine learning just statistics? \| Charles Isbell and Michael Littman and Lex Fridman
Lasting	:	4.38
Date of publication	:
Views	:	26 rb

Frames Is machine learning just statistics? | Charles Isbell and Michael Littman and Lex Fridman

Description Is machine learning just statistics? | Charles Isbell and Michael Littman and Lex Fridman

Comments Is machine learning just statistics? | Charles Isbell and Michael Littman and Lex Fridman

"what is the missing of the picture?"br"all the rest of it"brlmao
Comment from : whatthehead

The question "Is machine learning just statistics?", is more or less a YES There's a lot of statistics in machine learning, you really cannot do machine learning without statistics But this question also seems to implicitly imply that advancements in ML are not novel and not as valuable since "it's just statistics" This is definitely not true For if that was the case can we not just argue that even Statistics should not be considered valuable as it is just Mathematics?
Comment from : David

This podcast should be called “the nerd podcast”
Comment from : Himanshu Deswal

I agree that machine learning is just a form of statistics I think computational statistics is a perfect way of putting it Basically all you're doing is using large datasets to hone a hypothesis That's exactly what statistics is
Comment from : DrKnowsMore

Was für eine grandiose Show
Comment from : Hans Bleuer

A great synthesis of these views lies in probabilistic programming It embeds the language of statistics in a programming language, which necessarily means that questions of software engineering are relevant to the analysis of the data
Comment from : Austin Garrett

MATHEMATICS looking at its offsprings and be like- "that's cute :) :) :) " - while Machine Learning, AI, Computational Statistics, Physics, Computer sci etc are fighting amongst themselves
Comment from : Kartik Kamboj

Learning Machine is computational statistics! Those who say it is not does not know the field of statistics! Statistics can roughly be divided in too major fields: 1) probability theory and inference and 2) modelling, estimation and classification! Everything ML does is statistics empowered by powerful computers continuously fed by big data! The issue with the interpretation of dimensions higher than 4 is not a limitation of statistics (also not interpretable with ML), instead it is a limitation of human brain! Plus, coding in statistical software (SAS, SPSS, Minitab, etc and programming (Fortran, Pascal, etc) is a very part of the statistics field!
Comment from : Dina Cula

ML is definitely just a part of statistics
Comment from : Mwave

ML uses more non-parametric methods than traditional statistics Many Non-parametric methods are possible with higher computational power ML is just doing stats with higher computing power And ML is not about Software engineering or programming, software and latest computational methods are used in almost all subjects like computational biology or computational social science, which doesn't mean everyone is doing software engineering
Comment from : Faizan Mansoori

remove computable statisticswhats the result: nothing
Comment from : ivan avila

The guy on the right is saying that decisions in model design and stuff like the loss function you choose isn't a part of statistics? Is this satire? If anything, statistics is far bmore/b focused on the meta aspects of a quantitative analysis than ML, and has been since its inception
Comment from : Vainbow

Classical Lex question
Comment from : The Colony Engineer

Those who argue ML is not statistics, they don't know either
Comment from : logicverse

Yes Machine Learning is about rules, but those rules have statistical foundations, there are certain assumptions that need to be fulfilled in order to apply those algorithms Also, we have probabilistic classifiers in ML that rely heavily on probabilistic and statistical principles
Comment from : José González