Actually… 21 Sep 2009 … I know a lot of philosophers who can do math and physics (and some have PhDs in both physics and philosophy). See more comics from Calamities of Nature Humor Philosophy Science
Cognition My talk online 31 Oct 201931 Oct 2019 I recently gave a fuller version of my talk “Comprehension as Compression”, and I placed the slideshow on Slideshare. You may also download the PDF: Read More
Book Oh dear… 21 Sep 2009 Whenever a book is touted as a “paradigm shift”, it sends up a major warning flag: Put succinctly, what Alva Noë is offering in Out of Our Heads is nothing short of a paradigm shift, complete with an incisive criticism of the status quo of neurosciences and a suggestion for… Read More
History The best scientific ideas 20 Feb 2010 Over on talk.origins Roger Ebert’s review of The Creation, a film about Darwin’s private life, was mentioned, which starts out Darwin, it is generally agreed, had the most important idea in the history of science. Thinkers had been feeling their way toward it for decades, but it took Darwin to… Read More
I’m still waiting for someone to come up with bad philosophy based on the problem of P=?NP. It may just be that they are waiting for that to be solved before they do anything. However, the problem by nature has so many easy soundbite explanations that I’m quite confident it will be heavily abused once it is resolved. Soundbitedness seems to be the main criterion for attracting such abuse.
Philosophers couldn’t do that: it’s just too hard. Now, where did I put that Löwenheim/Skolem paradox? I’m here all week. Try the veal.
Actually… … I know a lot of philosophers who can do math and physics Yeah, but are they postmodern philosophers? As for P vs. NP, some might say that tying NP to human creativity (as some have done) is overreaching, philosophically speaking. I personally don’t think it’s overreaching, so it’s not bad philosophy to me. Misunderstanding NP has certainly led to bad physics, e.g. “nature can solve NP-complete problems in polynomial time”; “quantum computers can test an exponential number of possibilities simultaneously, thereby solving NP-complete problems in polynomial time”.
Lyle? Has anyone actually claimed that quantum computers can have solve NP complete problems in polynomial time? A quantum computer can solve factoring in polynomial time using Shore’s algorithm but that’s a much weaker claim. I would incidentally argue that tying NP to human creativity is massive overeach since the problem type being discussed are all very rote, not the sort of thing that creativity often comes into play.
Yes, people have made that claim. Do you read Scott Aaronson’s blog? Check out the saga of Geordie Rose and D-Wave. NP-hard problems are not rote by nature. The ones studied in CS theory courses tend to be rote, yes, but NP as a class includes many interesting natural problems. Take the problem of natural language processing: if we define it as finding a model of language that explains all existing human utterances, we can easily argue that it is NP-hard. The space of possible models is exponential in size, but any given model can be verified in polynomial time (in the size of the number of natural language sentences). The whole field, then, is about coming up with creative heuristics that do a good enough job of solving problems most of the time. Furthermore, every single human child (apart from some that are developmentally disabled) is able to learn some heuristic for this problem. Proving theorems is another (more classic) example of an NP-hard problem that we are able to apply creative heuristics to and solve some instances some of the time. If NP=P, there would certainly be no need for human mathematicians any more; mathematical creativity could be fully and efficiently automated.
Lyle, As I understand it Georgie Rose claimed merely that he could use D-Wave to get better approximate results for NP-complete problems. See for example: http://www.technologyreview.com/computing/18495/page3/ I disagree with your assesment of NP-hard problems. There are of course a number of different issues here. There are many problems which are NP-hard that are not NP-complete. For example, it isn’t hard to show that if one has a halting oracle then one can use it to solve NP problems in polynomial time. Thus, in one sense HALT is NP-hard even though it obviously is not in NP. If P=NP, human mathematicians would still be needed. Mathematical creativity is not limited to such problems but also to much more difficult problems that cannot be easily modeled as a single NP complete problem. Given a reasonably complicated axiomatic system “Is there a proof of Y of length l or less?” is in NP. “Is there a proof of Y?” is not in NP. Human ingenuity will thus still be helpful.