Amazing! (I can't find any other great word to describe my feeling). I have watched many linear algebra RUclips videos, and many of them were great. But this video cleared some of my doubts that other videos failed to do so. Keep going on! 👍
Especially that connection of cosets with covectors. And the explanation of why we don't care about positions of vectors. And, why do we fix vectors on origin in linear algebra.
wow, Im glad you mentioned MTB! Professor Grinfield just answers all your questions the moment they arise! he was the first youtuber that made me buy a book, his introduction to tensors :) that being said I dare say there is a small gap at basic linear algebra, so I'm glad youre making this series
I agree that MTB is an amazing channel, especially if you want to dive deep into the details. I'm glad that you appreciate our effort to close the gap at the basic level.
Dark mode. Nice. But white on light blue is a bit hard to read - the two most lit one's at 10:31 addition and multiplication panels. Dimmed colors is more readable.
Sorry for the shock 😆 Don't worry, soon enough we will talk about geometric algebra, and we will multiply vectors in so many different ways that it will be hard to keep up.
what a beautiful video with unique flow of the motives behind the concepts and how to do it. Really thanks for your efforts. Just wanted to ask are these animations by manim? Is there any simpler and faster way to do similar visuals?
Thank you for your appreciation. The animations are programmed in Python, using a custom library that is similar to Manim. I don't know of a faster/easier way to make animations. Maybe AI will help at some point ;-)
Non-linear isn't as bad as non-convex. Linear things are nice because knowledge about local properties becomes knowledge about global ones. For example, the derivative or slope of a flat line is the same everywhere. The next best situation is to have convex (aka concave) things because local knowledge at least provides global estimates. For example, once you know the local derivative/slope, you can figure out how the global one will behave even though it's not constant globally. Non-concave curvy things are not so nice to be able to extrapolate. This is something I learned regarding optimization but I think ideas generalize to other fields that classify things as linear or curved.
This is one of the most interesting comments we've had on the channel. It's easy to miss just how important concavity/convexity is. It got me thinking that perhaps the next-next-best-thing in the list would be monotonicity? Food for thought. Thank you!
@@AllAnglesMath That's flattering to hear! I heard a quote that was originally from R. Tyrrell Rockafellar: "The great watershed in optimization isn't between linearity and nonlinearity, but between convexity and non-convexity." And yes, it seems that monotonicity (at least in terms of the Hessian / second derivative) is the useful property because it implies that local min/maxima are also global. For functions like cubics or sinusoids, any one extremal point may not be global. For non-convex functions, some form of feasible set search is required such simulated annealing.
Awesome. Would love to know which tool you use for animations. Edit: a little doubt: in 28:35, if the basis is not orthogonal, the decomposition would not be the projection, I think that statement in the video is wrong or what am I missing?
Will you introduce basis (dual vector / covectors / linear forms / linear maps) as horizontal lists and coordinates (standard linear algebra vectors) as vertical lists?
Yes, that's the plan. We will write covectors as row vectors so that the product of a covector and a vector can be written as an ordinary matrix product.
@@AllAnglesMath Very interesting, glad for your answer! I'd like to watch a visual representation of this fascinating advanced topic, rarely brought into discussions. And I'd suggest eigenchris channel's tensor playlists. Maybe you put a link to his work, if you will.
@@linuxp00 Eigenchris is brilliant, and will be one of the main sources of inspiration for the series on tensor algebra. If you look carefully, you can even see his logo in one of our videos already.
honey wake up new All Angles video
Glad to see you back. I sincerely hope I never make a video that puts you back to sleep 🤣
Amazing! (I can't find any other great word to describe my feeling). I have watched many linear algebra RUclips videos, and many of them were great. But this video cleared some of my doubts that other videos failed to do so. Keep going on! 👍
Especially that connection of cosets with covectors. And the explanation of why we don't care about positions of vectors. And, why do we fix vectors on origin in linear algebra.
YAY! the linear algebra series
I am sure your channel will reach millions believe me I just find it and hit subscribe
Thank you, and welcome!
wow, Im glad you mentioned MTB! Professor Grinfield just answers all your questions the moment they arise!
he was the first youtuber that made me buy a book, his introduction to tensors :)
that being said I dare say there is a small gap at basic linear algebra, so I'm glad youre making this series
I agree that MTB is an amazing channel, especially if you want to dive deep into the details. I'm glad that you appreciate our effort to close the gap at the basic level.
Amazing 🤩
Wonderful!
Dark mode. Nice. But white on light blue is a bit hard to read - the two most lit one's at 10:31 addition and multiplication panels. Dimmed colors is more readable.
That's a very good point. I will try to keep it in mind in future videos. Thanks!
Geometric Algebra has warped my brain... when you said we don't want to multiply vectors I almost cried
Sorry for the shock 😆
Don't worry, soon enough we will talk about geometric algebra, and we will multiply vectors in so many different ways that it will be hard to keep up.
Simply fantastic to start with
what a beautiful video with unique flow of the motives behind the concepts and how to do it. Really thanks for your efforts.
Just wanted to ask are these animations by manim? Is there any simpler and faster way to do similar visuals?
Thank you for your appreciation.
The animations are programmed in Python, using a custom library that is similar to Manim. I don't know of a faster/easier way to make animations. Maybe AI will help at some point ;-)
And can you sharethe library you used? Thank you
Non-linear isn't as bad as non-convex. Linear things are nice because knowledge about local properties becomes knowledge about global ones. For example, the derivative or slope of a flat line is the same everywhere. The next best situation is to have convex (aka concave) things because local knowledge at least provides global estimates. For example, once you know the local derivative/slope, you can figure out how the global one will behave even though it's not constant globally. Non-concave curvy things are not so nice to be able to extrapolate.
This is something I learned regarding optimization but I think ideas generalize to other fields that classify things as linear or curved.
This is one of the most interesting comments we've had on the channel. It's easy to miss just how important concavity/convexity is. It got me thinking that perhaps the next-next-best-thing in the list would be monotonicity? Food for thought. Thank you!
@@AllAnglesMath That's flattering to hear! I heard a quote that was originally from R. Tyrrell Rockafellar: "The great watershed in optimization isn't between linearity and nonlinearity, but between convexity and non-convexity."
And yes, it seems that monotonicity (at least in terms of the Hessian / second derivative) is the useful property because it implies that local min/maxima are also global. For functions like cubics or sinusoids, any one extremal point may not be global. For non-convex functions, some form of feasible set search is required such simulated annealing.
Thank you!
Awesome. Would love to know which tool you use for animations.
Edit: a little doubt: in 28:35, if the basis is not orthogonal, the decomposition would not be the projection, I think that statement in the video is wrong or what am I missing?
Will you introduce basis (dual vector / covectors / linear forms / linear maps) as horizontal lists and coordinates (standard linear algebra vectors) as vertical lists?
Yes, that's the plan. We will write covectors as row vectors so that the product of a covector and a vector can be written as an ordinary matrix product.
@@AllAnglesMath Very interesting, glad for your answer! I'd like to watch a visual representation of this fascinating advanced topic, rarely brought into discussions. And I'd suggest eigenchris channel's tensor playlists. Maybe you put a link to his work, if you will.
@@linuxp00 Eigenchris is brilliant, and will be one of the main sources of inspiration for the series on tensor algebra. If you look carefully, you can even see his logo in one of our videos already.
Now I wish I knew how to bake a cheesecake...
Can you post more frequently?
They have the complete series already uploaded on their Patreon page, I believe