« first day (1080 days earlier)      last day (2549 days later) » 

 
3 hours later…
6:22 PM
@vzn What is $Sigma_2^p$? Can you explain what you think it is? My intuitive reading of this would imply P != NP, but if this were the case, the authors would have said so.
 
vzn
6:50 PM
@ThomasKlimpel hi whats new? its apparently the "polynomial hierarchy", not too familiar with it myself, it can be defined with "alternations", he seems to be trying to prove the noncollapse of the hierarchy at the 1st level, but am not clear if this also means from known theory P!=NP (youd think he would say that if so, dont see it). a little more on reddit. reddit.com/r/compsci/comments/4nvdcc/…
 
vzn
In computational complexity theory, the polynomial hierarchy (sometimes called the polynomial-time hierarchy) is a hierarchy of complexity classes that generalize the classes P, NP and co-NP to oracle machines. It is a resource-bounded counterpart to the arithmetical hierarchy and analytical hierarchy from mathematical logic. == Definitions == There are multiple equivalent definitions of the classes of the polynomial hierarchy. == Relations between classes in the polynomial hierarchy == The definitions imply the relations: Σ i ...
 
Well, the people at reddit also seem to think that it implies P != NP. And they further seem to think that it implies that this paper is crockery. However, not mentioning that it implies P != NP would actually reduce the likelyhood of unreasonable crockery a bit, because it indicates at least that the authors don't try to make their enormous claims too openly visible.
 
7:45 PM
@vzn The people at reddit are right, P=NP implies NP^NP=P. This paper claims to prove P != NP^NP, which is basically equivalent to proving P != NP. My opinion is that this paper is a very serious attempt. It won't be easy to disprove that paper. My guess is that it will be subject to normal peer review, which will determine in due time whether it is correct or not.
 
Polynomial hierarchy is taking the idea that NP is P but with a "there exists... Such that..." Quantified in front of it, now generalize for a chain of quantifies. Eg there exists a graph, such that for all nodes, there is some other node, ((some statement )), hence alternations between existential and universal quantifiers
Only formally defined with oracles
That's the best way of thinking about PH in my opinion. Obviously it all comes down to rigorous Turing machines in the end, and you can never fully generalize
Anyways that's where the idea of an "alternating" machine comes in
 
However, the name Norbert Blum is probably not the famous person you expect, but just a "normal" professor.
 
vzn
8:10 PM
@ThomasKlimpel (oops)
yes thx @Kurt maybe the PH is roughly the same as counting alternations in QBFs? QBF solution is in Pspace and iiuc PH=Pspace is known
In computational complexity theory, a decision problem is PSPACE-complete if it can be solved using an amount of memory that is polynomial in the input length (polynomial space) and if every other problem that can be solved in polynomial space can be transformed to it in polynomial time. The problems that are PSPACE-complete can be thought of as the hardest problems in PSPACE, because a solution to any one such problem could easily be used to solve any other problem in PSPACE. The PSPACE-complete problems are widely suspected to be outside the more famous complexity classes P and NP, but that is...
o_O didnt notice that. the paper is from Feb of this year & seems to have attracted no notice except recently...? the author has some record/ other papers on arxiv in complexity theory but seems not the specific area of PH arxiv.org/find/cs/1/au:+Hauptmann_M/0/1/0/all/0/1
@ThomasKlimpel oracles can be tricky, not sure that P!=NP^NP is the same as P!=NP. intuitively maybe, but whats the proof? this reminds me of the distinction between cook vs karp reductions which is an active area of research.
 
9:03 PM
@vzn The proof is P = NP -> NP^NP = P and P != NP -> P < NP <= NP^NP -> P != NP^NP. This shows that P != NP is equivalent to P != NP^NP.
I wouldn't worry about the fact that the paper is from Feb. It will go through the normal peer review channels, and the author will retract it in case serious gaps are found.
 
vzn
10:07 PM
@ThomasKlimpel dont think there is really such a thing as a "normal peer review channel" for the "very big" open problems & eg the ("relatively recent") perelman proof attests to that. while admittedly circumstantial its not a good sign that its gotten no notice among "heavyweights" for ~¼ yr.
 
@vzn You submit a paper to a journal (or a conference), and the editor takes care to sent out your paper to appropriate peers able to judge it quality. The only reason why this would not happen if is the author simply didn't submit his paper to a journal, or if no journal was willing to accept the paper for peer review.
 
vzn
@ThomasKlimpel am aware of "normal peer review process". there is no indication so far hes submitted it anywhere (other than arxiv). also "peer review" has multiple meanings, one very specific in scientific contexts, and there are others less formal...
 
I looked at his homepage, and all his older papers were published in journals (in addition to arXiv versions), which seems to indicate that he has the habit to submit his papers to journals. The gap between the dates of the arXiv versions and the dates of the journal versions also gives you some indication of how long it takes until a journal paper appear (at least for him).
 
vzn
did you hear of the deolalikar proof attempt from a few yrs back?
@ThomasKlimpel "big proofs" (not just in pages, but attacking forefront open problems) can take very long to peer review. years. am making no assumptions that he submitted it anywhere until hearing otherwise.
 
Of course. But here Deolalikar is to blame, for not properly reacting to the feedback he got.
 
vzn
10:18 PM
@ThomasKlimpel he is to "blame" for a wrong proof. & maybe not communicating openly after (others) discovered/ isolated it & brought it to his attn. ie (afaik) "never"
 
The failure to communicate properly is much worse than a wrong proof or wrong paper.
 
vzn
@ThomasKlimpel lol, the paper itself is a communication. dont really know what pt youre trying to make here.
think deolalikar deserves some respect for his attempt. as like others, really dont like his lack of response to its detected failure but dont think its something to dwell on either.
 
If you want to "blame" somebody for a wrong paper, then you only get into a situation were nobody will ever admit that his paper was wrong. But what really annoys other scientists is if they spend the time to read your work and give you feedback, and you just dismiss it.
 
vzn
@ThomasKlimpel (afaik) he didnt dismiss it, he was just silent/ unresponsive, think there may be some small )( difference there.
wrong proofs are "annoying" to everyone, incl esp the author & reviewers etc...
anyway, one pov wrt deolalikar (& why was bringing it up) is that its roughly the 1st case of a sizeable cyberspatial collaborative "peer review" that crossed blogs & other mostly cyber media (email etc) in near-real-time. ie, a new (so far somewhat unrecognized) form of peer review. the perelman proof history also generally supports this idea.
 
Maybe I have read too many unfinished papers, and gave feedback to too many different people. I have no problems with bad English, faulty proofs, incomplete descriptions. But I have problems when the typical expectations about interactions are ignored.
 
vzn
10:35 PM
@ThomasKlimpel yes, think you are much more generous/ patient than typical or at least elite researchers in complexity theory who have no patience for amateur attempts, and therefore atypical in that regard.
tend to lean/ follow/ encourage your stated philosophy/ approach myself.
have a few blogs wrt peer review expressing similar ideas etc
 
@vzn Most computer programs have some bugs. Most papers have some bugs too. The paper with some wrong statements is rather the rule than the exception. The question is how you react when somebody points out a serious problem. For software, you typically use a bug tracker, and maybe you fix some of the reported bugs in future versions.
 
vzn
@ThomasKlimpel yes, exactly, have advocated the same concept of "buggy papers" in much the same way, but it is not a widespread/ accepted paradigm/ metaphor among mathematicians/ computer scientists for misc reasons.
"proofs are like code that runs in logicians [mathematicians|comp scientists|etc] heads"
wrt your criteria, Hauptmann is on rev3, and it seems to be only this latest rev that drew attn by a single user on reddit who posted it, and now an apparently small audience eg reddit commenters, incl ourselves.
 
I think the typical mathematician acknowledges that there are many bugs in papers, but he claims that most of those bugs can be fixed easily, just like most bugs in computer programs can be fixed easily. And I guess this assumption is actually a good model of what is really going on.
 
vzn
(btw, as/ like you wrote, do not regard hauptmann as an "amateur attempt"...)
@ThomasKlimpel yes and automated thm proving work does suggest that these assumptions while maybe generally ok are not so straightfwd sometimes, and those researchers are causing some reevaluation of logic/ rigor in mathematics.
the CFSG is a great "recent" case study, found great some links on that... think there is maybe some ATP work on it iirc...
 
In mathematics, the Feit–Thompson theorem, or odd order theorem, states that every finite group of odd order is solvable. It was proved by Walter Feit and John Griggs Thompson (1962, 1963). == History == William Burnside (1911, p. 503 note M) conjectured that every nonabelian finite simple group has even order. Richard Brauer (1957) suggested using the centralizers of involutions of simple groups as the basis for the classification of finite simple groups, as the Brauer–Fowler theorem shows that there are only a finite number of finite simple groups with given centralizer of an involution. A group...
 
vzn
10:47 PM
yeah exactly F-T was just subject to ATP at MS, getting some attn, it was quite (somewhat unexpectedly?) nontrivial.
 
Why would you assume that Hauptmann has not submitted his paper to a journal? If he wants to claim the 1M dollar, then the original rules say that he has to do that anyway. Those rules were not followed for Perelman, but he didn't claim the 1M dollar in the end either.
 
vzn
@ThomasKlimpel sigh. said am not assuming he submitted it which as you might realize from logic is not the same as am assuming he did not submit it :P
 
Ah, I see
 
vzn
think maybe 1 us needs to email him asap to ask him that & other questions, dont you think? :) :P
maybe he is more "responsive" than deolalikar o_O
another key/ obvious/ glaring question for him is why he didnt put P!=NP in the title or abstract or even contents...
it does seem already something is off :(
but on other hand, the same could be said of perelman! ... except it wasnt his claims/ results, lol!
 
11:03 PM
Yes, but I would wait now before emailing him, because he will get some feedback caused by that reddit link. In a month or two, he might already have a better idea whether the feedback was related to the content of his paper, whether anybody has really tried to read it, and whether some people confirmed that his proof might work.
 
vzn
@ThomasKlimpel he may not be aware of the reddit response at all so far. whatever, will probably email him fairly soon if nobody else does & no other info/ stuff emerges soon. proofs ("attempts") like this only come along once every few years, its worth an email at least, could type one out in less time spent so far in this chat :)
 
Well, there you go...
 
vzn
(further googling) anyway Hauptmann seems to have much better credentials than Deolalikar so one can be hopeful...
have you heard of the Berman-Hartmanis conjecture? always found it fascinating (but it is somewhat downplayed by the pros, maybe increasing my interest). he has a paper on it...
he is credited as editor of Approximation Algorithms for NP-Hard Problems (2004) by Kannan and has his own submission there
 
No, I didn't really heard of it. I now looked it up at wikipedia. Does it basically conjecture that all NP-complete languages are paddable?
 
vzn
@ThomasKlimpel it is maybe related to padding, not really that sure myself! :| ... but the idea of a very strong symmetry among NP complete problems and inside the class is quite compelling, it seems to be shown already to a large degree in other ways... and also note it would be a proof of P!=NP if true... am always looking for/ like to collect alternative formulations of P=?NP problem...
anyway you are saying P != NP -> P != NP^NP (ok!) but the converse is not known is it? wonder what is known
 
11:28 PM
The converse P = NP -> P = NP^NP is known as well.
 
vzn
meant by the "converse", P != NP^NP -> P != NP? think that is not known is it?
 
The sentences "A -> B" and "not B -> not A" are logically equivalent.
 
vzn
11:43 PM
ok. am having trouble wrapping my brain around this. just never heard anyone suggest that P!=NP would be proved by "noncollapse" of PH. guess it makes perfect sense...?
 
I'm going to sleep now...
 

« first day (1080 days earlier)      last day (2549 days later) »