Excellent video. When "s" appears alone in Hungarian, it's often pronounced as a voiceless palato-alveolar fricative ("sh"). When "sz" appear together, as in Lovász's name, the noise they make is just "s" (as used in English). This is the reverse of the Polish, and may be confusing. "Low voss".
I'll try! In the Gram Schmidt process we do b_k = b_k - u_{k,j} b_j for 0 < j < k. Intuitively, this is what makes the basis orthogonal, as we are subtracting out the projections of the basis vectors onto each other, leaving behind on the perpendicular components of b_k in the directions of each b_j. However, in LLL we are working in a lattice, so we can't guarantee orthogonality. That is, b_k = b_k - u_{k,j} b_j is probably not in our lattice. For it to be, we would need u_{k,j} to be an integer. So in the LLL alg, we do the same thing as in the gram schmidt process, but we round the u_{k,j} term to the nearest integer. This produces a basis that is "orthogonal enough" while still remaining in the lattice
At 7:51 , the power of 2 should be placed at mu_[1,0] instead of outside the bracket right?
Yes, great catch! I would update that, but I'd have to rerender the entire video
@@stevenschaefer7314 Anyway it's a great and helpful video! Thank you!
Spent a couple hours to find resources on how these lattice reduction algorithms work. At last, found this video! Great job !!!
EXCELLENT!
Nice work! Was just looking for a explanation of this algorithm to implement an Integer relation algorithm
Thank you so much!
This is awesome, thank you!
Great video, nice slides!
How to do the algebraic number approximation in Matlab or Python or any other tools ?
Excellent video. When "s" appears alone in Hungarian, it's often pronounced as a voiceless palato-alveolar fricative ("sh"). When "sz" appear together, as in Lovász's name, the noise they make is just "s" (as used in English). This is the reverse of the Polish, and may be confusing. "Low voss".
Extraordinary!
Great video!
12:51 mu_[1,0] = 1/3 and the rhs should be 0.52? Anyways this is a very well made and helpful video, thank you :)
Please if can you help me I need a program algorithm about my design and evaluate cryptosystem based on braid groups
Thank you for this nice video! This was really helpful! Is this done with manim?
Thank you! Yes this was made in manim
This clearly has that 3b1b look, did you create it with his python library?
Yes, I used manim to make this!
Hello! Could you please explain what the purpose of the line:
bk = bk - [uk,j]bj
is please?
I'll try!
In the Gram Schmidt process we do b_k = b_k - u_{k,j} b_j for 0 < j < k. Intuitively, this is what makes the basis orthogonal, as we are subtracting out the projections of the basis vectors onto each other, leaving behind on the perpendicular components of b_k in the directions of each b_j.
However, in LLL we are working in a lattice, so we can't guarantee orthogonality. That is, b_k = b_k - u_{k,j} b_j is probably not in our lattice. For it to be, we would need u_{k,j} to be an integer. So in the LLL alg, we do the same thing as in the gram schmidt process, but we round the u_{k,j} term to the nearest integer. This produces a basis that is "orthogonal enough" while still remaining in the lattice
Can you put link of jeff Suzuki video of explaining Lavash condition?
ruclips.net/video/XEMEiBcwSKc/видео.html
How you get b2 * this
I don't understand the question
1:00
L