This is simply outstanding. Have been following your videos for about 2 months. Really interested in SINDy and This AI/ML+Physics Series. Don't think there is anything of this level available on UA-cam. Thank You for this brilliance.
im surprised to not see fourier neural operators or neural operators considered in this series. long time fan though and im firmly in ml now because of one of your videos on ml for turbulence closure (back then i was a cfd guy). I still have really liked this series so far its thorough as always.
Great video! I recently learned that you are a professor at UW (where I am wrapping up my undergraduate and about to start my PhD). I have been doing research in applied ML for photonics systems for some time now and it’s awesome to see UW represented at large! Maybe we will cross paths at some time.
It would be great if you could cover scalability issues and how you overcome them in a ML setting where by design you have a ton of data. Constrained optimization, even convex, is very hard.
As someone who works in Game Engines on real-time simulation this entire playlist has been my evening routine for the week. I haven't been this excited about a lecture series since I graduated uni. The games industry has gotten really good at rigid body dynamics (although real-time still has accuracy issues for complex mechanisms), but everything beyond that is a complete approximation. Most cloth physics simply acts as a post-processing step that cannot affect the rigid body dynamics. Similar state in fluid simulation as well. Trying to bind all three into a single cohesive system is prohibitively expensive because to get each system to affect the other becomes unscalable using traditional solvers. I feel like I am seeing the light with these lectures, fascinating! Quick question, in these lectures you are referencing code examples for PINN and SINDy systems. I haven't seen these in descriptions yet, are they coming or should I look for the source somewhere else? I ended up purchasing the Data Driven Science and Engineering book and saw some overlap with the topics discussed here, so I was wondering if some of this content is on that same github page?
Hi Dr. Brunton! I will be going to school for mechanical engineering but I love the concepts that you explain in your videos. Any advice on what classes to take or what to minor in? Thanks for uploading these vidoes. I also am reading your newest book- or at least trying to :)
Is there a difference in the goals of SINDy vs Symbolic Regression? I understand they work differently, but do they do the same thing, or am I missing something?
![AI/ML+Physics: Recap and Summary [Physics Informed Machine Learning]](http://i.ytimg.com/vi/GCz6afDVy5Y/mqdefault.jpg)
![AI/ML+Physics: Recap and Summary [Physics Informed Machine Learning]](/img/tr.png)
![Physics Informed Neural Networks (PINNs) [Physics Informed Machine Learning]](http://i.ytimg.com/vi/-zrY7P2dVC4/mqdefault.jpg)
I don't remember being as excited for a video to drop as I am with this series, thank you steve bry the fluid machine learning guy
This is simply outstanding.
Have been following your videos for about 2 months. Really interested in SINDy and This AI/ML+Physics Series. Don't think there is anything of this level available on UA-cam.
Thank You for this brilliance.
im surprised to not see fourier neural operators or neural operators considered in this series. long time fan though and im firmly in ml now because of one of your videos on ml for turbulence closure (back then i was a cfd guy). I still have really liked this series so far its thorough as always.
Don't worry, neural operators will be coming up in a few weeks. Glad you like the series!
These presentations/visualizations are among the best I've seen. Thank you!
The Physics-Informed Dynamic Mode Decomposition video: ua-cam.com/video/lx-msllg1kU/v-deo.html in case anyone else is looking for it.
Where are all the links to other videos promised in the video? Thanks again for the series!
Really high quality videos, looking forward to the rest of the series. Looking to create a digital twin at work!
Great video! I recently learned that you are a professor at UW (where I am wrapping up my undergraduate and about to start my PhD). I have been doing research in applied ML for photonics systems for some time now and it’s awesome to see UW represented at large! Maybe we will cross paths at some time.
Are there missing links which was mentioned during lectures. I can’t find the links to get more extended videos.
It would be great if you could cover scalability issues and how you overcome them in a ML setting where by design you have a ton of data. Constrained optimization, even convex, is very hard.
As someone who works in Game Engines on real-time simulation this entire playlist has been my evening routine for the week. I haven't been this excited about a lecture series since I graduated uni.
The games industry has gotten really good at rigid body dynamics (although real-time still has accuracy issues for complex mechanisms), but everything beyond that is a complete approximation. Most cloth physics simply acts as a post-processing step that cannot affect the rigid body dynamics. Similar state in fluid simulation as well. Trying to bind all three into a single cohesive system is prohibitively expensive because to get each system to affect the other becomes unscalable using traditional solvers. I feel like I am seeing the light with these lectures, fascinating!
Quick question, in these lectures you are referencing code examples for PINN and SINDy systems. I haven't seen these in descriptions yet, are they coming or should I look for the source somewhere else? I ended up purchasing the Data Driven Science and Engineering book and saw some overlap with the topics discussed here, so I was wondering if some of this content is on that same github page?
Thanks you professor, your video is truely helpful, I am trying to dig into the content your talked about, but can't find corresponding links
@steve! Amazing works and gifts!! Many thanks!! Do you have any practical examples to test?
Hi Dr. Brunton! I will be going to school for mechanical engineering but I love the concepts that you explain in your videos. Any advice on what classes to take or what to minor in? Thanks for uploading these vidoes. I also am reading your newest book- or at least trying to :)
what textbook?
Is there a difference in the goals of SINDy vs Symbolic Regression? I understand they work differently, but do they do the same thing, or am I missing something?
Any implementations so far?
I'm very interested in applying this to my work - "physics based synthetic data"
Amazing series! Part 3 is somehow missing though!!
Amazing serie
Thank you sir.
Thank you!
amazing
Look like to me Sindy is Taylor series expansion from data
I find this path too demanding, generally.
It's an understatement that whole Ph.D.s can be dedicated to solving an optimization problem.