| personal webpage | https://homepages.cwi.nl/~ducas/ |
| github | https://github.com/lducas/ |
Even better, they even provide an easy-to-run #algorithm (#Matlab code)! Let's look at Cauchy—Schwarz in the pictures below. This is a very simple example, but it can do much more. This should tell you enough to make you want to try it!
Code (and paper proving its correctness) available on Greg Valiant's page: https://theory.stanford.edu/~valiant/pap_link.html
[I am working with a student to convert it to other languages and improve the interface]
2/2
Reposting here, as this could be of interest to the #mathematics community: the automatic inequality prover of Greg and Paul Valiant, which lets you to automatically check and (dis)prove a whole bunch of inequalities.
Basically any inequality that's a combination of Hölder and ℓp-monotonicity-type #math #maths #tools
You give a candidate inequality in that form, and the output will be either a #proof that it doesn't hold, or a message saying that it does& how to prove it!
1/2
I ran in the following head- 
scratcher:
f(n,m) can be computed at a cost C either from f(n-1,m) or f(n,m-1). f(0,0) is given.
I don't know the exact input values in advance, but I do know that calls to f will happen with decreasing m and increasing n. I'll make at most k calls, with 0 <= m,n <= k.
If I'm greedy (taking the shortest path from memoized calls), that can cost me up to O(C * k^2).
Can I do better ?
My impression on Mastodon:
Until recently:
- niche geek / academic experiment, can't compare with twitter. Can't even discover new profiles easily! 🤦♂️
Now:
- wow, Mastodon is not trying to steal my attention in any way. No ads. No junk content inserted on my timeline. This is invaluable 🤗🥲😍🤟
Clément Canonne
Nils Weisensee
nsi_xyz
eprint
seinecle