https://arxiv.org/abs/1708.03486 I realize there are dozens

Norbert is quite good. The paper is 38 pages long... I'll read it tomorrow at work - but it seems legit.

I just finished reading it and while there were some parts I couldnt follow it looked actually really good.

>>9105212
He has a few papers on data structure lower bounds and a paper on his pet model of Boolean network functions published in a low-impact journal. He is not a reputable researcher on the topic, and the paper makes a bizarre claim that a lower bound method for monotone circuits extends easily to general boolean circuits. I don't, for example, see how Theorem 6 couldn't be applied to the majority function to extend the known Ω(n2) lower bound for monotone circuits to general circuits, contradicting the known O(log n) constructions for computing majority with general circuits.

About a year ago, there was a similar submission: https://arxiv.org/abs/1602.04781
which also claims P!=NP (in different words). If you look at the end of the paper, the author thanks Norbert Blum for reviewing (and indeed google confirms they are both professors at the University of Bonn). Not to say that Professor Blum's submission is invalid, but it certainly hurts credibility in my mind.

It's also worth noting that Blum's paper has some of the original criticisms I had of Hauptmann's, namely that it doesn't seem to connect to existing research on the topic, such as why this approach succeeded where others failed or how this approach relates to the massive amount of open questions that are closely tied to the P vs NP problem.

To be specific, the paper I linked claims that the Polynomial Hierarchy does not collapse all the way, but a well known result is that if P = NP then PH collapses entirely, therefore making any claims to the contrary is also making a claim against P vs NP.

Why a picture of RCT?

>>9106894
Because it's a great game.

>>9105752
You also posted this on plebbit.

>>9107040
Well, he's not wrong.

>>9105752

How can that guy just publish a paper that claims to solve the P vs NP problem but also reviews a paper that has claimed to have solved the same problem 1 year before?

If people decide Blum's proof is correct, will there be a fight in the university of Bonn for who gets the 1 million dollars?

>>9107107
Yes, a fight for death. It's a custom in german university culture.

[spoiler]If you're serious: the old proof was incorrect, so he had his chance[/spoiler]

There is a discussion about this on the CS Theory Stack Exchange

https://cstheory.stackexchange.com/a/38811

In particular, Karl Wimmer seems to take issue with the assertions being made in Theorem 6, which was also criticized in >>9105752

"If f is monotone and computed at g0, it is fine to take DNF(g0), apply absorption and the R operator, and the resulting DNF′(g0) represents ff. Using this "one-shot" construction, Theorem 5 is fine--on to Theorem 6. I glossed over this definition of DNF′(g0)

What I can't see is why the gate-by-gate apply-absorption-and-RR-as-you-go construction of DNF′(g0) on pages 27-28 does the same thing. This seems necessary for the gate-by-gate analysis in Theorem 6 to work, unless error from this construction is accounted for. I mean, not every function can even be represented by a DNF with terms with only non-negated or negated literals, but for each node g, DNF′(g) seems to always have this form. What if there is a node g in my network such that res(g) has no such representation?"

If there is any problem with this proof, it seems that Theorem 6 is where it would be.

File: IMG_6954.jpg (687KB, 2936x1240px) Image search: [Google]

687KB, 2936x1240px

understanding math formulas

i consider myself one standard deviation more intelligent than the avg person (and i could name a few interdisciplinary accomplishments to prove that)

yet when i decided to get into machine learning - and not just learn some software packages but actually understand the concepts behind it - i quickly realized that i totally suck at math.

i mean i flip through a book on ML and it's full of this kind of stuff and i have not a CLUE IN HELL what it means.

what would be the best way to get started to understand mathematical notations?

mind you, i understand the operators, like the brackes, the dividing line, what the E sum means

but i do not understand the actual meaning of the formula

and those books are full of them and dont explain anything (not blaming them, i know it's my fault not to understand)


---
so basically my question is:
do i just have to focus my ass and keep reading books on ML and at some point it will make click and ill understand

OR

do i need to start at some far more basic, unrelated topic (sth like advanced calculus idk)?

---

would there be any quick-to-consume literature i could check out? i tried searching my uni's library for "uni math for beginners" or "math formulas for beginners" but didnt find anything

thanks in advance