So you can solve a multi dimensional optimization problem by finding the global minima for individual dimensions and then "freeze" them, go along the next variable? I wonder if that can optimize a neural network... Will try to find the literature
A piece of well-intentioned feedback: it's cool that you use the manim library for visualizarions, but you don't need to rip off 3blue1brown's voice inflections too! I'm sure it's unintentional, but your narration sounds like you are imitating him (with the rising tones and mid-sentence pauses).
Amazing animation and explanation! At first, it was difficult to understand the narration due to pronunciation, but I got used to it quickly 😆 Great job on colab with Polylog as well 👏👏👏
hello! great video, I was looking into your repo since I would like to integrate this presentation in a general intro to Mathematical Optimization. Unfortunately I think asset are missing. Any chance you can commit them. or if you have a ppt from this it would be perfect (I was planning to use manim-pptx)
The jawdropping moment when you open the source code of the video and find out that it is more than 11 THOUSAND lines!! How long does it take to make this video??
Hey Tom, brilliant video! What I don't exactly understand is what you mean with "ratio", when you say th e "we choose the largest non-positive ratio". A ratio for me is relation/proportion. Ans seemingly, you are comparing the negatives of the corresponding constants (-4000 and -5000). Could you elucidate where I am mistaken?
Yeah that animation is really not showing what's happening well, sorry. By ratio (say for the 2nd inequality), we take the constant value (4000) and divide it by the coefficient the variable we're loosening (-1), i.e. 4000/-1=-4000. For 3rd inequality, we do the same: (5000/-1=-5000). If the coefficient for x_2 were -2, we would do the same: (4000/-2=-2000). Hope this helps!
Sorry in what parallel universe does the professor: A) look at your homework And B) not claim your new discovery for themself I am not falling for it, not gonna do my homework.
구글코리아 개발자 직원들이 고객의 스마트폰을 해킹하고 동영상및 광고를 개발자입맛데로 마음데로 조정하며 고객의 스마트폰을 자신들 장난감처럼 갖고 논다. 내부 고객 프라이버시 지침은 있으나마나고 완전 개꼴통들 집합소다. 중요한 얘기할땐 휴대폰 끄고 아주 사적인 내용은 휴대폰에 저장금지 내말 믿어 구글코리아 개발자놈들이 몰래 다 처듣고 처본다. 특정 아이디 지정해서 어떤 동영상을 보여줄건지 어떤 광고를 보여줄건지도 세밀하게 지들 입맛데로 조절할수있다. 내말믿어 구글코리아 개발자놈들 아주 음흉한 놈들이다. 중요한 얘기할땐 휴대폰끄고 아주 사적인 내용은 저장금지. 일부 부정직하고 생각없는 구글코리아 개발자 직원들의 내 스마트폰에 해킹및 계정에대한 여러가지 조작행위로 인해 구글과 유튜브의 시작부터 지금까지 구글팬이었던 나를 안티 구글로 만들고있다. 난 진심으로 저 개발자들 개인적으로 만났으면 좋겠다. 인터넷뒤에서 증거있냐 식으로 그만 깝치고
My college professor was not bad, but this video is something else. Really utilising the technology to present complex topics in such an amazing way. I am aware of how complex and immensly time consuming these types of videos get, but please do continue making these.
9:44 non-positive ratio? Didn't you mean non-negative ratio? ... Understand, which value would be the s_x when in other side of equation, and ignoring other s_x