Thread by @EchoStatements

THREAD: There are lots of great educational and entertaining machine learning YouTube channels out there. @AndrewM_Webb and I have compiled this list to share some of the best channels that we’ve found.

@AndrewM_Webb Crash Course AI (@TheCrashCourse) serves a gentle introduction to machine learning, presented in an informal style. Almost no math, but lots of motivating examples, intuitive visualisations and historical background.

@AndrewM_Webb @TheCrashCourse 3Blue1Brown’s Neural Networks playlist (@3blue1brown) Shows from first principles how neural networks work, with great visualizations and clear explanations, building up the viewers intuitive understanding and introducing the underlying mathematics

@AndrewM_Webb @TheCrashCourse @3blue1brown @3blue1brown also has great series introducing linear algebra and calculus, both vitally important for anyone wanting to develop a thorough understanding of machine learning topics

@AndrewM_Webb @TheCrashCourse @3blue1brown Arxiv Insights: On his channel, Xander (@xsteenbrugge) uploads deep-dives into popular ML topics, often focusing on one or more recent research papers. His videos don’t shy away from technical details, but are clear and well-written.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge As well as being the host of Crash Course AI, Jabril (@jabrils_) makes his own videos, increasingly focused on AI/ML in application. His videos are silly and irreverent, but with a lot of work behind them. Always fun to watch.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ StatQuest (@joshuastarmer) focuses on stats rather than ML, but has gentle introductions to a lot of machine learning models and concepts, with clear and well-explained examples.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex has a lot of great educational ML vids, mostly focused on coding/implementation. But to me, his most interesting videos are his series on training a self-driving car in GTAV: a great example of the tenacity required to apply ML to hard problems.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex Lex Fridman (@lexfridman) manages to get an impressive stream of prominent researchers to sit down for long, insightful interviews. Also available as a podcast, the channel gives a glimpse into the world of ML research.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman Two Minute Papers (@karoly_zsolnai) gives brief summaries of recent research papers. Often focusing on visually appealing research, the channel showcases some of the more interesting things happening in the world of ML, as well as other fields.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai Although no longer active, DeepLearning.TV (@deeplearningtv) produced some great short videos on ML topics, with surprisingly high production value.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv Similarly giant_neural_network (@joncomo) has some good introductory tutorials on neural networks, but hasn’t posted ML related content for a while.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo If you want to get into machine learning in a more systematic way, there are also some great free online courses. As a disclaimer, neither @AndrewM_Webb or I have taken these courses, but they look to us like they’d be a great place to start learning.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg’s @coursera course is a free 56 hour intro to ML, building to neural networks from linear and logistic regression, and covering kernel methods, unsupervised learning, and applications. Includes programming assignments as well as lectures coursera.org/learn/machine-…

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera fast.ai's free courses aim to make deep learning widely accessible. For example, "Practical Deep Learning for Coders": course.fast.ai covers backprop, image classification, language processing, GANs, cloud deployment, and more. @jeremyphoward @math_rachel

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel For an even *more* formal introduction to machine learning, here are some good textbooks, three of which are freely available online.

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel The "deep learning bible", covers i) deep learning as it is practised and ii) current research topics. Reading this book will give a student enough context to understand many current research papers. Freely available online: deeplearningbook.org
@goodfellow_ian

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian Bishop’s book is great for those starting out in ML. Needs some familiarity with calculus, probability, and linear algebra, but builds up to more difficult concepts gradually while still managing to rigorously cover a lot of ground. (@ChrisBishopMSFT)
microsoft.com/en-us/research…

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT With a huge variety of ML topics, this might be more of a reference book than one you'd read cover to cover. It's comprehensive and self contained, with introductory chapters on probability, Bayesian statistics, and frequentist statistics. Intro chapter: cs.ubc.ca/~murphyk/MLboo…

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT If you’re interested in the mathematical framework behind ML, I’d recommend “Foundations of Machine Learning” by Mohri et al. While the material covered is denser than the previous books, it is well presented and well motivated. (@atalwalkar)
cs.nyu.edu/~mohri/mlbook/

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar ADDENDUM: Since publishing this thread, we’ve had some great suggestions. Here are a few things that we missed the first time round:

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar Henry AI Labs (@CShorten30) has some really nice videos discussing recent research papers. A relatively small channel that deserves to be bigger! (thanks to @Inoryy for the suggestion)

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar @CShorten30 @Inoryy Brandon Rohrer’s (@_brohrer_) is another channel we didn’t find in our original pass, but it looks like he's making some great stuff. I've only watched his backprop video so far, but will definitely watch more!

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar @CShorten30 @Inoryy @_brohrer_ Stanford’s CS231n: Designed by @drfeifei and @karpathy, this deep learning/computer vision course has a fantastic reputation. Course materials are freely available online (thanks to @dunchen_master7, @AhmadMustafaAn1 & @sreekar339 for the suggestion).
youtube.com/playlist?list=…

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar @CShorten30 @Inoryy @_brohrer_ @drfeifei @karpathy @dunchen_master7 @AhmadMustafaAn1 @sreekar339 PyImageSearch (@PyImageSearch). While I don’t know enough about the paid courses and books to recommend them personally, there are loads of interesting detailed articles on the free-to-access blog. Worth checking out!(Thanks to @gpakosz for the suggestion) pyimagesearch.com

@AndrewM_Webb @TheCrashCourse @3blue1brown @xsteenbrugge @jabrils_ @joshuastarmer @Sentdex @lexfridman @karoly_zsolnai @deeplearningtv @JonComo @AndrewYNg @coursera @jeremyphoward @math_rachel @goodfellow_ian @ChrisBishopMSFT @atalwalkar @CShorten30 @Inoryy @_brohrer_ @drfeifei @karpathy @dunchen_master7 @AhmadMustafaAn1 @sreekar339 @PyImageSearch @gpakosz While on the subject of blogs, I’m a huge fan of Distill (@distillpub) . With a focus on clarity of presentation, Distill takes cutting edge ML research papers and transforms them into amazing interactive articles @shancarter @ch402 @arvindsatya1 distill.pub