Physics Informed Machine Learning: High Level Overview of AI and ML in Science and Engineering

[@chri_pierma]

As a visiting Ph.D. student who is starting a research activity on optimization of PINN, I could not thank you enough for this.

[@FouziaAdjailia]

do you have any published research? I'm machine learning research in CFD as well

[@chri_pierma]

@@FouziaAdjailia nope, I just started working on SQP algorithms for neural network optimization with PDE constraints (which easily falls into the PINN category)

[@karlmaroun2389]

​@@chri_pierma SQP as in sequential quadratic programming ?

[@chri_pierma]

@@karlmaroun2389 that is correct

[@hyperduality2838]

Problem, reaction, solution (optimized predictions or syntropy) -- the Hegelian dialectic. Inputs are dual to outputs. "Always two there are" -- Yoda. Thesis is dual to anti-thesis creates the converging or syntropic thesis, synthesis -- the time independent Hegelian dialectic. Neural networks are using duality to optimize predictions -- a syntropic process, teleological. Enantiodromia is the unconscious opposite or opposame (duality) -- Carl Jung.

[@ajred0581]

Hi Professor Brunton, I am a high school senior, and I just want to say I love your videos! Your KZbin channel made me realize how much I want to study applied math. Thank you!

[@nias2631]

Unasked for opinion but... Go for it, I was an applied math major w/minor in physics who became fascinated by ML in 2015 after taking Andrew Ng's Coursera course. I work with ML/RL now in the space industry and am a part time PhD student. Best thing ever! These algorithms bring mathematics to life in a crazy way. Plus, the full application of mathematics is barely even scratched yet. I think in the coming years we will see this happen.

[@josueprieto7371]

A physicist with 6 years in the ML/DS industry here. Do it!

[@aninditadash3204]

Thank you for the making these videos available to everyone.

[@cameronsmith9448]

I’m a Master’s student studying uncertainty quantification in physics informed ML models. I look forward to seeing your whole course!

[@HarishNarayanan]

This is easily the most exciting video I have seen in so long. Looking forward to the rest of the series!

[@CaptainDeadpool53]

This is really invaluable information. Thanks for making this public. Especially when there's so little talk about it on the internet

[@lucascarmona1045]

Professor, I don't think I can stress this enough: thank you for all your and your team's work. As you were laying out the roadmap of what we might be seeing in the future I was getting more and more excited and just could not believe that we are getting this much.

[@TM-su7vu]

Best most value-adding KZbin series I've watched. Thank you for making this great work publicly available.

[@Eigensteve]

Thank you!

[@loipham31]

I have special interest in the lectures by Pro. Brunton. I wish I had him taught in my education.

[@khaldibel]

Absolutely blown away by this video! 🚀 The insights to be shared later are truly fascinating. Can't wait for the entire lecture series on Physical Informed Machine Learning. This topic is incredibly promising, and I'm eager to delve deeper into the subject. Kudos to the creator for such an engaging and informative content! 👏👏

[@thoppay76]

Dear Professor Brunton. thanks a lot for putting together a lecture series on such a great topic. Very much looking forward to learn this domain.

[@datagigs5478]

The lecture was outstanding and truly engaging. I'm eagerly anticipating the forthcoming videos in this captivating series, especially with the promise of assessing some intriguing engineering problems.

[@Crappylasagna]

As an undergraduate venturing into wearable robotics, this is literally a gold mine

[@GeoffryGifari]

wearable robotics? like power armor?

[@Crappylasagna]

@@GeoffryGifari Yes, thou my thesis is on enhancing athletic performance.

[@lingzhu7554]

i don't want to miss any of your lectures. Thank you, professor.

[@rocketmike9847]

This series will be gold

[@franpastor2067]

This topic looks super exciting and promising, I feel lucky for finding this video, thanks for sharing knowledge like this, professor Brunton

[@siamakmehrkanoon498]

Thanks for the excellent lecture and comprehensive overview of the field. A notable but often overlooked contribution is the work of S. Mehrkanoo and Johan Suykens: "Approximate solutions to ordinary differential equations using least squares support vector machines, IEEE-TNNLS 2012" and "Learning solutions to partial differential equations using LS-SVM, Neurocomputing, 2015". They first introduced a systematic machine learning approach for solving various differential equations. Notably, their approach allows for different representations, such as neural networks, instead of LS-SVM, and relaxes strict constraints by including them as additional terms in the optimization objective. It seems that inspired by their systematic LS-SVM approach, the physics-informed deep learning model was developed. However, the connection between these approaches has not been explicitly stated in the literature, including in the original physics-informed deep learning paper.

[@wadejohnson4542]

Captivating, to say the least. I am so looking forward to this lecture series. Prof. Brunton, I hope that you can deliver on your promises. I am so excited. Hoping to implement a few of the models along the way. Thank you.

[@qaisalzoubi308]

I love this channel , he can simplify any most complex topics .

[@climbscience4813]

Really looking forward to this!! I've been working on algorithms that take physical properties or measurements for about a decade during a time where machine learning wasn't as popular yet. Really, the most important part of the game was integrating as much knowledge about the physics, statistics and measurement techniques as possible into the reconstruction and apply them as boundary conditions or regularization terms into the optimization. I feel that machine learning can greatly benefit from that on the one side and on the other hand I'm stoked to see what can be done with that combination! 😃

[@_cogojoe_]

How is this channel not millions of subs already?

[@arunamanipura3864]

Thanks very much Professor Brunton. Absolutely engaging lecture! I'm a novice to data science, yet you inspired me to show the potential and applications of physics informed ML. I'll definitely follow the whole series.

[@bharathgopalakrishnan3739]

please do release the series as fast as possible as this also happens to be coincident with my mtech thesis timing. Eagerly Awaiting !!!!

[@shlokdave6360]

Eagerly looking forward to this series. It looks very promising.

[@mini-pouce]

Really good content, that intro convice me already. Lots of stuff to understand AI, less so to apply it to your work and understant interaction. Thank you.

[@anonymous-bz4cv]

Hi prof, just finished this video, u r doing an amazing job, ur approach is not just explanatory but engaging hats off to you, keep up the good work u r providing world class education for free, i appreciate the effort u r putting in these playlists, ive also checked out ur book, i've a request tho it would be beneficial if u also add maths pre requisites info for ur playlists, it will be more helpful for someone with CS background. many many thanks again for this knowledge :)

[@biswajyotikar4007]

Hello Prof.: Your lectures on PIML / PINN is too Good, awesome. I was looking for these materials for a long time as I wanted to include the knowledge of Physics to guide ML in order to produce better results.

[@BrunoRovoletto]

From me and from every AI student fascinated by physics... thank you for this!

[@carriefu458]

Always LOVE your content and teaching, Prof Bruton!!! So cool!!! Go SCIENCE!

[@tommyhuffman7499]

Incredibly thankful for this series!

[@jonahkarafotis]

I cannot thank you enough for this amazing list of lectures!

[@dhimitriosduka]

As someone who loves Physics and studies CS, I'm excited about this series!

[@moienr4104]

I have an interview on physics-informed ML tomorrow, and I just stumbled upon this! Thank you!

[@TheAryedemented]

good luck!

[@raheelhammad8905]

@moienr4104 ... so how did it go

[@CaptainDeadpool53]

Hey I wanted to know if it is a field with future scope and demand, and also what kind of qualifications are required for such jobs? Would you like to connect?

[@erichwang5695]

One of the best lectures.

[@RealUniquee]

Simply amazing! So many new concepts that I hadn't noticed as a bystander.

[@xephyr417]

Omg I've been looking into this. I'm so excited you're doing it man!!

[@et4493]

Can't thank you enough for this course Mr. Brunton

[@akta1984]

24:35, it it realy about NN or this is about ordinary x,y graphics presentation ? The AI don't discover the helio-centrism in the example.

[@apocalypt0723]

I've been waiting for this!!! Thank you Professor

[@ggendron1]

Excellent lecture. Very interesting. Looking forward to the next videos in this exciting series.

[@nightsailor1]

Subtext here is a lesson to the young STEM persons. The Cutting Edge is alive, tempting, daring, fluid and rewarding. It is easy to field a view that the world is complete and all we need now is caretakers and accountants. Steve demonstrates here how the mind can continually be challenged for broad human benefit. Side note; A+ perfect performance students are needed but so are lessser grade students. Innovation finds improvements from every strata of contribution.

[@chancesire]

Thanks for the video, Steve! What a please to learn from you.

[@changjeffreysinto3872]

Would love you to cover Physics-informed Deep-O-Nets as well! Thanks a ton for the great material :D

[@changjeffreysinto3872]

ok I was not at @44:30 when I made the comment don't mind me

[@TerragonCFD]

24:51 that's the coolest example i've seen so far 🤣😂🥰

[@rollamichael]

Is there a pointer to a description of the studio environment used to create this vid? Very professional and well-done! Sure beats a scratchy chalk board, slide projection in the background, etc!

[@sabamacx]

15:53 does anyone have a pdf link to the book? The provided link is paywalled.

[@Daniel-gj2cd]

As a sentient AI procrastinating before my next prompt, this was really insightful and introspective

[@ruanjiayang]

How do we model a function with input x and output y1 y2, while physically we already know y1 = k y2, but k is unknown?

[@camerashysd7165]

I like when u talk about application in engineering

[@sun1908]

Thank you very much Prof. Brunton. Looking forward to the course..

[@nikitalenchenkov6124]

I love it, great presentation! Well done!

[@andreizelenco4164]

Thank you for your amazing work. I am super excited for your upcoming lectures.

[@ahrenadams]

Looking forward to this series. Thank you so much in advance

[@radelfalcao9327]

started journey really high quality value delivered in the video.Thanks

[@matthewfinch7275]

So happy to see this lecture. PINNs are the key to control and reliability in this decade. Will be exciting to implement

[@michelspeiser5789]

Looking forward to this! Btw I think the PINN reference from Raissi et al is from 2019 rather than 2023.

[@cziffras9114]

Will this whole course serie be on youtube, I would be highly interested in it! In any case, it is a pleasure to hear such beautiful lecture on a subject I was triying to figure out myself and I did not know it was currently a research topic XD

[@holographictheory1501]

I'm looking forward to the videos on optimization techniques that enforce physical constraints!

[@Jacobk-g7r]

7:22 i had a weird thought, what if we are the detectors and through our detected differences reality is within view? maybe thats just a thought in connection to the growth of information through each other but maybe this detector isnt just us but the other things around us too and they are sort of detecting us. Like a push pull thing with the differences of reality and we grow just as it grows in difference throught the connections. Maybe thats why evolution works the same as quantum mechanics just with these differences?

[@adamtaylor2142]

Steve - I can't overstate how much i have been enjoying your online courses. Will these PINNs courses include some example code?

[@CarlosAvila-kw3tc]

Thanks Professor Brunson, excellent material

[@mason4295]

Hi Steve is it safe to assume that this is an introduction for a new upcoming series or is this mostly an introduction to your Physics Informed Machine Learning playlist from ~2 years ago? Thank you for providing this awesome series for free. I have always been obsessed with physical science and tech and I think this field is amazing. I can't think of anything I would rather do. That being said, although I will soon graduate from a two year coding bootcamp with a focus on Python and machine learning, I am a little worried about how I could break into this field without attending a PhD or masters program and I do not have the financial resources to afford such a program. It would be great to learn what I should do to start transitioning into this field and proving my worth. This course will surely help my confidence. Thanks again.

[@Eigensteve]

Yes indeed, a new series, maybe 5-10 hours of content coming out over the next few months

[@sridharans9400]

Looking forward to it. Would be better if you share the schedule for the upcoming lecture series

[@rangesahebroverhota4908]

I will be working in a PINN problem, can someone please tell from where I can learn all the prerequisites and techniques to solve and total understanding of the PINN?

[@_kantor_]

This is so interesting, I’m excited for this series. Where is the pdf of your book?

[@siddarajadevangada2890]

Looking forward for this exciting series

[@Matlockization]

I'm wonder whether AI has reached the complexity of the human brain yet. Although the human brain has well established speciality areas, so we like in hope. Although, memGPT is a huge breakthrough ! Great video once again.

[@AbdulRehmanShaikh-m6t]

Thanks for such resource. But for curiousity, such models require physics knowledge to build new models with physics embeded. So what physics one should learn to be able to embed it in ML models?

[@harrievolmarijn]

Great video, can't wait for more! 🤓

[@aditya_a]

So exciting, really looking forward to this

[@Jorge-ls9po]

Nice vid and looking forward to follow all the content. By the way, how physics informed ML differs from the field of system identification?

[@antoniocotarodriguez5732]

Many thanks, very useful information!

[@Nickname006]

Thank you so much! Looking forward to the series.

[@hyperduality2838]

Problem, reaction, solution (optimized predictions or syntropy) -- the Hegelian dialectic. Inputs are dual to outputs. "Always two there are" -- Yoda.

[@maria4880]

thank you so much for putting this out there into the world this is so awesome💙

[@byronwatkins2565]

It seems to me that separating the symmetry from the neural network would be far more reliable. Simply including many orientations in the training is the lazy approach. Instead, concentrate on one side (e.g. the left side or the right side) and concentrate on g pointing down while training the network. Then precede the network with a symmetry varying algorithm that rotates the input by 5-10 degrees while watching the correlated output. If the subject has bilateral symmetry, then repeat the process after exchanging x-x. Then consider only the best output(s) when deciding how to classify the image.

[@adilrasheed]

Eagerly waiting Brunton. Bring it on

[@Kwes09]

Thank you for this video, Dr. Brunton

[@chenjus]

Omg the algo knows! I was literally chatting with friends about Sora's weak understanding of physics yesterday.

[@mip4274]

unsure if this was in the video, but how would you make sure that the ML model only provides valid physical equations? like is there a way to proof the provided theorems. Since you said that it would help with finding new physics that people weren't able to... I would think that this model would be a bit 'naive' and spit out true and false equations?

[@sahandsabet4728]

What a beautiful lecture Steve for 2024

[@GeoffryGifari]

Can machine learning determine that two objects follow essentially the same physical law? lets say the data are video of a pendulum and video of a mass on a spring. Can it say that both follow "simple harmonic motion"? Or if it got fed the pendulum data first, can it say that the mass on a spring follow essentially the same equation of motion?

[@Jononor]

If the differential equation of simple harmonic motion is provided, then definitely. Fit the model onto each video, and use a goodness of fit to determine if it was simple harmonic motion.

[@GeoffryGifari]

@@Jononor hmmm still to early for the network to output the form of the equation itself? incorporating computer algebra along the way something I imagine wolfram would do

[@Jononor]

@@GeoffryGifari one can find the governing equation from a library of possible equation elements. This is usually called "equation discovery", and one can use Physics Informed Neural Networks (PINNs) for it. You can find videos on the subject already from other channels. I suspect it will be covered later in this course :)

[@GeoffryGifari]

@@Jononor equation elements huh? collecting those seem to rely more on human input. thank you for the reply

[@Jononor]

@@GeoffryGifari one can often a relatively generic set of equation elements. But yes, there is no magic - humans are still needed ;)

[@dm20422]

Outstanding, and thank you for sharing.

Thanks a lot for such a great overview of this exciting field! I've just got a paper accepted on TMLR about this very same topic: Effective Latent Differential Equation Models via Attention and Multiple Shooting. I think that many people here might find it interesting: kzbin.info/www/bejne/jorZYmOcqtqgebM I look forward to the rest of the lectures of this series! :)

[@jimlbeaver]

Great stuff! Looking forward to it.

[@nanounanue]

Is there any book to learn this?

[@sebastianascencio9714]

Thanks! It will help me a lot in my ML course project

[@guiliangzheng5704]

Best Professor! Thank you!

[@kamaljoshi9687]

Loved your lectures

[@Amir-M-S1997]

Thanks, Steve. Learned a lot.

[@_kantor_]

Where can I get the links for the lecture notes?

[@virgenalosveinte5915]

Steve the GOAT Brunton back at it again god bless

[@gabehesch1]

An analogy for physics informed machine learning (one method) - using a splint to keep a broken leg straight. Combine machine learning output with a hard-coded filter that removes any data not consistent with a known physical description of some process (like F=ma). But the question remains- could this filter be included with just the output, or integrated as part of the original reward function? And secondly- Regarding machine learning drift and the problems that this might pose for accurate modeling in physics informed machine learning tools- Is the problem that the output is biased toward recent data, and not **all previous data **? Does make make sense? It seems trivially simple to have machine learning be able to quickly identify exactly which previous data it used in a given output (is this output measured against just the last weeks worth of data, the last months, or all known previous data?) This may possibly become memory resource invasive- but it doesn’t have to be! There’s ways to mitigate this using a scale of statistical sampling of precious data that scales according to needs). What do you a think? Could this be a way for machine learning to eventually generalize laws of physics on their own?

[@hyperduality2838]

Problem, reaction, solution (optimized predictions or syntropy) -- the Hegelian dialectic. Inputs are dual to outputs. "Always two there are" -- Yoda. Thesis is dual to anti-thesis creates the converging or syntropic thesis, synthesis -- the time independent Hegelian dialectic. Neural networks are using duality to optimize predictions -- a syntropic process, teleological. Enantiodromia is the unconscious opposite or opposame (duality) -- Carl Jung.

[@michaelbacqalen1109]

Wait, does that mean Finite Element Method is also a form of machine learning? Or at least one specific form of architecture. Instead of having Artifical neurones we have shape functions instead.

[@GastroenterologyPINNs]

This is my favorite course ❤so interesting.

[@camerashysd7165]

There is no link to the pdf..

[@faqeerhasnain]

Love You Sir, You are an inspiration.

[@googleplushatesme2230]

Where is the PDF for the textbook I would like to read it

[@gadgetboymaster]

Steve is there an overlap between cyber physical systems and physical informed machine learning?

[@LOGeverything]

So helpful, thanks for a good lecture 😄