Lynx Roundup, June 24th 2018 Daily roundup of Data Science news around the industry, 6/24/2018. Matthew Alhonte 24 Jun 2018 What is the role of qualitative methods in addressing issues of replicability, reproducibility, and rigor? "Let me count the ways..." A few weeks I read an article in Physics Today which prompted me to revise and resubmit an old post I cobbled together in response to the BBC television series Wonders of the Universe in which I argued that the title of that programme suggests that the Universe is wonder-ful, or even, in a word which has cropped up in the series a few times, `awesome’. When you think about it the Universe is not really `awesome at all’. In fact it’s extremely overrated. Take this thing, for example: This is an example of a galaxy (the Andromeda Nebula, M31, to be precise). We live in a similar object. Of course it looks quite pretty on the surface but, when you look at it with a physicist’s eye, such a galaxy is really not as great as it’s cracked up to be, as I shall now explain. We live in a relatively crowded part of our galaxy on a small planet orbiting a fairly insignificant star called the Sun. Now you’ve got me started on the Sun. I know it supplies the Earth with all its energy, but it does the job pretty badly, all things considered because the Sun only radiates a fraction of a milliwatt per kilogram. By comparison a human being radiates more than one watt per kilogram. Pound for pound, that’s more than a thousand times as much energy as a star. So, in reality, stars are bloated, wasteful, inefficient and not even slightly awesome. They’re only noticeable because they’re big. And we all know that size shouldn’t really matter. In short, stars are extremely overrated. But even in what purports to be an interesting neighbourhood of our Galaxy, the nearest star is 4.5 light years from the Sun. To get that in perspective, imagine the Sun is the size of a golfball. On the same scale, where is the nearest star? The answer to that will probably surprise you, as it does my students when I give this example in lectures. The answer is, in fact, on the order of a thousand kilometres away. That’s the distance from Cardiff to, say, Munich. What a dull landscape our Galaxy possesses. In between one little golf ball in Wales and another one in Germany there’s nothing of any interest at all, just a featureless incomprehensible void not worthy of the most perfunctory second thought. So galaxies aren’t dazzlingly beautiful jewels of the heavens. They’re flimsy, insubstantial things more like the cheap tat you can find on QVC. What’s worse is that they’re also full of a grubby mixture of soot and dust. Indeed, some are so filthy that you can hardly see any stars at all. Somebody needs to give the Universe a good clean. I suppose you just can’t get the help these days. And then to the Physics Today piece I mentioned at the start of this article. I quote: Star formation is stupendously inefficient. Take the Milky Way. Our galaxy contains about a billion solar masses of fresh gas available to form stars—and yet it produces only one solar mass of new stars a year. Hopeless! What a waste of space a galaxy is! As well as being oversized, vacuous and rather dirty, one is also pretty useless at making the very things it is supposed to be good at! What galaxies clearly need is some sort of a productivity drive or perhaps a complete redesign using more efficient technology. So stars are overrated and galaxies are overrated, but surely the Universe as a whole is impressive? No. Think about the Big Bang. Well, I don’t need to go on about that because I’ve already posted about it. Suffice to say that the Big Bang wasn’t anywhere near as Big as you’ve been led to believe: the volume was between about 115 and 120 decibels. Quite loud, to be sure, but many rock concerts are louder. To be honest it’s a bit of an anti-climax. If I’d been in charge (and given sufficient funding) I would have put on something much more spectacular. In any case the Big Bang happened a very long time ago. Since then the Universe has been expanding, the space between galaxies getting emptier and emptier so there’s now less than one atom per cubic metre, and cooling down to the point where its temperature is lower than three degrees above absolute zero. The Universe is a cold, desolate and very empty place, lit by a few feeble stars and warmed only by the fading glow of the heat left over from when it was all so much younger and more exciting. Here and there amid the cosmic void a few galaxies are dotted about, like cheap and rather tatty ornaments. It’s as if we inhabit a shabby downmarket retirement home, warmed only by the feeble radiation given off by a puny electric fire as we occupy ourselves as best we can until Armageddon comes. In my opinion the Universe would have worked out better had it been entirely empty, instead of being contaminated with such detritus. I agree with Tennessee Williams: BRICK: “Well, they say nature hates a vacuum, Big Daddy. BIG DADDY: “That’s what they say, but sometimes I think that a vacuum is a hell of a lot better than some of the stuff that nature replaces it with.” So no, the Universe isn’t wonderful. Not at all. In fact, it’s basically a bit rubbish. Again, it’s only superficially impressive because it’s quite large, and it doesn’t do to be impressed by things just because they are large. That would be vulgar. Digression: I just remembered a story about a loudmouthed Texan who owned a big ranch and who was visiting the English countryside on holiday. Chatting to locals in the village pub he boasted that it took him several days to drive around his ranch. A farmer replied “Yes. I used to have a car like that.” Ultimately, however, the fact is that whatever we think about the Universe and how badly constructed it it, we’re stuck with it. Just like the trains, the government and the weather. There’s nothing we can do about it, so we might as grin and bear it. It’s being so cheerful that helps keep me going. Follow @telescoper Debugging with intelligence via probabilistic inference | the morning paperthe morning paper a random walk through Computer Science research, by Adrian ColyerAdrian ColyerStatistical Significance Tests for Comparing Machine Learning Algorithms Roundup Matthew Alhonte @MattAlhonte Supervillain in somebody's action hero movie. Experienced a radioactive freak accident at a young age which rendered him part-snake and strangely adept at Python. Hackers and Slackers Newsletter Join the newsletter to receive the latest updates in your inbox. Your email address Subscribe Please check your inbox and click the link to confirm your subscription. Please enter a valid email address! An error occurred, please try again later.