I get confused trying to learn Gaussian Processes | Learn with me!
Вставка
- Опубліковано 12 чер 2024
- Watch me stutter for 2.5 hours in the uncut video: / 47543982
View the recap doc here: docs.google.com/document/d/1t...
▶️ More videos: / intuitiveml
New Twitter! / intuitive_ml
____________________
Learn more:
The main article in the video: distill.pub/2019/visual-explo...
Wiki: en.wikipedia.org/wiki/Gaussia...
Other sources used throughout the video:
Cornell: www.cs.cornell.edu/courses/cs...
Katherine Bailey: katbailey.github.io/post/gaus...
Peter Roelants: peterroelants.github.io/posts...
0:00 Intro
1:15 Predictions Visualized
24:20 SKIP HERE IF ATTN. SPAN == SQUIRREL
#IntuitiveML, #machinelearning, #LearnWithMe
I never realized just how loud my keyboard was and how often I say "sure", "yeah", "okay", "umm", or "right" until I edited this video.
Remember to support: www.patreon.com/intuitiveml
This sentence contains exactly threeee erors.
Sorry this didn't get posted last month, I lost two weeks of time to procrastination.
This illustrate perfectly the saying: If you have no doubt, you din't understand ANYTHING. This is so confusing I didn't even get to the point of having doubts yet, the guy in the video, in the other hand, by having so many doubts, actually understood a lot already
You need more subs. The slower, casual style of explanation is lovely. GPs can get confusing. Great work.
Thank you very much. Yeah, I figured it would work better for GPs because I have gotten confused trying to learn them several times before, so I figured the normal short and sweet wouldn't work.
Very concise and deeper , it is the best .
What i like about your videos is that you give a personal insight of your thought process. That actually helps the observer of your videos to learn how to think for him self. Thats a great way to teach someone. more of these videos bro
This is a great format for learning! I just learned GPs in 30 mins (plus a few rewinds!) Also picked up some good study habits :)
This is a really interesting style of video. Gaussian processes have had me stumped for a while, the math in every single online guide seems completely unrelated to machine learning and is way too complicated anyway, but i seem to be learning something from you. Thanks!
The way you learn is inspiring to me! Thanks for the video!
@IntuitiveML =)) "Skip here if attention span == squirrel" is what describes my ADHD
At first I was like "what kind of video is that?" but that helped me a lot! thank you :)
I'm glad it helped you!
great video. Thanks for thinking out loud and talking us through it.
Please keep these coming!
.
I think you have a very good voice for doing these videos, an excellent understanding of the material (obviously), and the quick/dirty visual style really works (IMO). I'd love to keep coming back to your channel for quick videos that cut through all the minutia, so I can really get the core concepts in my head quickly. After that, there are hundreds of other videos available where I can go deep.
.
At least, that's what I like about your channel and what I'm hoping to get out of it. This exact video was fun to watch, seeing you get over the mental humps, but may I suggest you either stick with the real short videos or maybe pair each of them with a longer one like this video. (you could pair them in Playlists..2 at a time and/or color-code them, obviously with similar titles). That way if I watch one of your shorts (say blue, for fast), and I still don't get it, then I can turn to one like this (say green, for the long mental discovery), which walks me through the learning barriers (as if we are hiking together through the material and both face the same hills). Come to think of it, you'd cover both areas which are hard to find on UA-cam (the long hand-holding as well as the super-short concept-oriented methods). That would be perfect!
.
Anyway, it's just a suggestion.
What an incredible video. Nice work, I appreciate it
Thank you so much.
thanks, this video helped me a lot
Very nice video and style. Thanks! Shouldn't the interval against a single point be prediction interval and not the confidence interval?
Thank you
what a creative content !
Thank you for making this. Great stuff and entertaining.
I'm really glad you liked it! I was honestly pretty afraid that people wouldn't like this one because it's much longer and slower-paced than my usual content... 😅
Oh, boy, I had the same experience trying to learn about Malliavin Calculus (note the "trying" word is still there). 😂
From which website are you reading?
Distill: distill.pub/2019/visual-exploration-gaussian-processes/
Really great video, but I think there is some confusion over what is a Gaussian Process versus what is Gaussian Process Regression (which is what you are talking about for the most part). Anyways, great content :)
Fair enough! Thank you.
Nicely done