Bayesian logic applied to an old bromide

July 10, 2013 • 6:51 am

From today’s xkcd.

seashell

Is there any better series of science-related cartoons?

That said, there’s one factor left out here, which is the probability that you can hear the ocean given that you’re near the ocean. I suspect that’s high, but you could always be deaf, or the ocean could be quiescent. I’m sure readers will think of other problems. . .

34 thoughts on “Bayesian logic applied to an old bromide”

  2. I hear the ocean all the time even though I live in MN. It’s called tinnitus…one of the side effects of hearing loss.

      1. I know that. It’s the susurration of blood flow echoing in the skull cavities.

    2. Hey, still learning. Also in MN, also have tinnitus from hearing loss. Sucks, doesn’t it? There are times when my tinnitus is so loud, I couldn’t hear the ocean no matter how near I was to it. It’s like having the sound of an ocean in one’s skull. Come to think of it, that’s a hell of a “design flaw,” isn’t it?

      1. …whereas mine sounds like a high-pitched squeal (which I guess goes with the fact that my auditory deficit is in the higher frequencies).

      2. Since we came out of the ocean millions of years ago, maybe it’s our ancestral song.:)

    4. After all the comments, I’m convinced it’s a MN thing. BTW, I can’t decide if I like the high-pitched June bug or the idling truck better.

  4. Good comic, but it got Bayes’ theorem wrong. It should be

    P(O|S)=[P(S|O)P(O)]/P(S);

    the posterior probability is proportional to the likelihood times the prior probability (not divided by the prior probability).

    2. Great catch!*

      To internalize this concept I find it helps when one of the events is very low probability:

      O = P(someone’s an Ornithologist)
      S = P(someone’s Single)

      then:

      (small #1) = (middling #1) * (small #2) / (middling #2)

      * Humanity really needs an anti-confirmation bias pill.

  6. This reminds me of a piece that I installed for a wealthy collector back when I dabbled in the world of fine art.

    It was a neon piece that came unassembled, a number of neon tubes, mounting pins, transformers, bulk high voltage cable, that was installed directly on a wall. It consisted of the phrase “I listen to the ocean(s?)(roar?) and all I hear is you.”

    To create the piece the artist hand wrote the phrase on a piece of note paper and sent it to a neon craftsman to create it in neon. This piece was in the low six figure range.

    I have no wish to impose my aesthetic sensibilities on others, but while that might pass for bad art, it sure as hell is not fine art, in my book. The artist certainly didn’t put any serious creative effort into it. This is common in the world of fine art though. Even at the highest levels 90%, well 75% anyway, is crap. Marketing is king, just like selling used cars or getting a politician elected.

    2. These days art is all about concept and not really about execution.

      And, per Sturgeon’s Law, 90% of concepts out there are crap.

      If you ask me, this is bass ackward. Artists are feted for having sometimes clever, usually very general and nebulous ideas. The quality of the piece intended to represent this idea doesn’t seem to matter. I think concept is the easy part. Refining the concept until it manifests as a skillfully executed, compelling piece of art is where the artist proves herself.

      1. I couldn’t agree more. I have grown to despise the “concept is all” school of thought. The best art is, of course, art that exhibits both excellent concept and excellent craftsmanship / execution. Only a small percentage of concepts will be truly intriguing, and when it looks like a 1st grade group art project, that not being the artist’s intent, why should anyone have any respect for whatever concept the art was intended to relate? Art is a multifaceted endeavor. To concentrate on one facet to the exclusion of the others makes no sense to me. A turd with a little shiny spot on it is still a turd.

        What I found surprising after gaining some experience with the high end of the fine art world is that even at the high end Sturgeon’s Law still holds. I don’t base this only on my subjective aesthetic response to the art (though I acknowledge that could bias me), but on technical aspects as well. If you go through ArtBasel the majority of the art is poorly executed and the concept is obvious, cliche, shallow or just plain boorish. Of course the good stuff is truly amazing.

      2. Amen, brother.

        Concepts are a dime a dozen. Developing a concept and seeing it all the way through is what’s hard. The greatest artists are the ones who’ve mastered their craft

        Cage and Warhol each had some very important concepts about the ubiquity of beauty. What they said needed to be said — and, make no mistrake, each said it superbly. Each was very skilled at the craft; they just turned that craft on the banal — and the world is a much richer place because they did.

        But there’s really not much to be said about the banal, and they both said all that needs to be said. We don’t need anybody else to say the same thing…and, yet, so much “art” is people doing exactly that. Yeah, we heard that already — brilliant the first time, not all that interesting the second, and bloody annoying the fifteenth brazilian time.

        Cheers,

        b&

        1. It’s interesting that you bring up Cage. I’ve always thought of 4’33” as one of the quintessential “triumphs” of concept over content/execution (along with other notables like Ligeti’s Poème Symphonique or Kosugi’s Music for a Revolution).

          You make an good point that certain concepts, while worthwhile, don’t require extensive treatment, which is a big part of why much of the work that’s derivative of Cage et al is not worthwhile. I don’t think I’ve ever formally thought about the issue that way. But I also think that even skeptics tend to shelve their skepticism when faced with art. A lot of things get canonized that I think shouldn’t. Do we really need 4’33” to demonstrate that we can find aural beauty (or at least stimulation) all around us? Is that an idea for which someone should be celebrated?

          Complicating matters is absolute music. What is the concept that a Bach fugue is intended to represent? Much of the greatest music ever written isn’t tethered to, doesn’t spring from, any extra-musical concept. It’s simply the logic of how best to proceed from pitch to pitch, in purely musical terms.

          I guess I ultimately must concede that it’s only my preference to do so, but I give much more weight to content than to concept, across the board, always.

          1. I certainly understand where you’re coming from.

            My only point is that 4’33” is the clearest, most emphatic way I’ve ever encountered to viscerally demonstrate that our lives are filled with sound. And that five minutes is just about the perfect amount of time for that lesson to sink in.

            It’s something worth composing — once. And it’s something worth experiencing as a member of the audience — once.

            Once Cage did it, though, there wasn’t any remaining need for anybody else to do it…and, yet, they do.

            It’s much like Glass. His Four Organs was something worthy of a graduate student composing, maybe even as a small part of a thesis project. That so much of the rest of his compositions are barely indistinguishable from it is the mind-numbing tragedy. I’ll give him a bit of credit: he has a couple of interesting musical ideas every now and again. It’s just that an entire opera of his would best be distilled into a five-minute etude.

            Cheers,

            b&

          2. Re Glass: Agree.

            Most of the minimalist composers bring to mind that old quip (some say it originates with Stravinsky) about Vivaldi: he didn’t write 500 concertos; he wrote 1 concerto 500 times.

            🙂

          3. Reminds me of perhaps my favorite knock-knock joke.

            Knock, knock. Who’s there? Knock, knock. Who’s there? Knock, knock. Who’s there? Knock, knock. Who’s there? Knock, knock. Who’s there? Knock, knock. Who’s there? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip. Phillip who? Philip Glass. Wha…that’s not funny. Philip Glass. Oh…go fuck yourself.

            Cheers,

            b&

  9. I rarely hear the ocean when I pick up a seashell. We still call them seashells around the Great Lakes.

  10. That’s Bayesian Logic? Sorry for my ignorance but I never really looked into it before and I often heard it mentioned. Creationist use it a means to ‘prove’ god exist.

    In hope of clearing my ignorance some more is this really used in any real world way? Or is it a form of speculation prior to performing tests?

    1. Bayes’ Theorem is an uncontroversial result in probability theory. It’s use requires that you assign a probability to something being true prior to conducting your investigation. During the 20th century, few scientists thought you could assign such probabilities except in a few unusual cases, thus Bayes’ theorem had few uses for scientists. In the 21st century, more and more scientists seem willing to make these a priori assignments, and thus Bayesian reasoning is undergoing a renaissance (it had been popular in the early 19th century). For a critical analysis, see this paper by Elliott Sober, and for what’s up lately, this paper by Bradley Efron.

      GCM

      2. Yes, Bayes theorem is utterly uncontroversial, but Bayesian statistical inference was relatively little used until about 25 years ago.

        The big breakthrough was when statisticians (notably Adrian Smith) discovered Markov Chain Monte Carlo methods in the statistical physics literature, and realised that these gave the computational tools to fit far more detailed and realistic Bayesian models than ever before.

        Hence, Bayesian methods suddenly became practically useful.

        The problem of course is that often you don’t have prior information, but the approach still requires you to quantify your prior information even if you don’t actually have any.

Comments are closed.