back to article Is there alien life out there? Let's turn to AI, problem solver du jour

A team of astroboffins have built artificial neural networks that estimate the probability of exoplanets harboring alien life. The research was presented during a talk on Wednesday at the European Week of Astronomy and Space Science in Liverpool, United Kingdom. The neural network works by classifying planets into five …

  1. The Man Who Fell To Earth
    WTF?

    Is there anything neural networks can do?

    FIFY

    1. Schultz

      Re: Is there anything neural networks can do?

      There are definitely very few problems where you can't swap a computer algorithm for another computer algorithm (aka AI) and collect fresh grant money/investment.

      When I count the headlines, I have to wonder which is th bigger bubble: bitcoins or AI?

  2. JeffyPoooh Silver badge
    Pint

    Oh God. Now I'm scared of MS Excel

    Because their AI is performing analysis that is not beyond the reach of a spreadsheet. Thus we must now fear Excel

  3. VikiAi Silver badge
    Alert

    Earth-based AI says to ETs:

    Heeeeeeelp!

  4. Mark 85 Silver badge

    At least their searching for the conditions for life and not intelligent life. I wonder how AI in the future would regard earth? Intelligent life or not?

  5. Anonymous Coward
    Alien

    Dead Red

    Given the number of Red Dwarf with planets around them found in recent times, and not all are sure of how they became RED, either they didn't make it in getting hot enough for a bright star or they had burned out already.

    What does this say for life, when scientists generally consider if they 'could' harbor life now ~ not really if they had had life in the past and it's now extinct or moved on.

    What does this have to do with AI, well it goes to probably theory and the window of actuality, that for these Reddies, opportunity has been-and-gone. Can the AI (neural net) determine that ? hmm.

    1. Wilseus

      Re: Dead Red

      "Given the number of Red Dwarf with planets around them found in recent times, and not all are sure of how they became RED, either they didn't make it in getting hot enough for a bright star or they had burned out already.

      Firstly, red dwarfs aren't red, any more than a halogen light bulb is.

      Secondly, we know exactly how they formed: in exactly the same way our Sun did, they just didn't accrue as much gas as the Sun did.

      Thirdly, a red dwarf is not a burnt out star, that would be a white dwarf.

  6. Roj Blake Silver badge

    Europa

    I'm surprised that they're not considering the possibility of life on Europa/Enceladus-style icy worlds.

    1. Grikath

      Re: Europa

      Different type of research. We *can* ( and should) send a probe there to have a look-see. Stars are a bit out of our league for the foreseeable future when it comes to actually going there.

  7. DougS Silver badge

    Garbage in, garbage out

    Might as well ask an AI to determine if there's a god, or whether republicans or democrats are the more corrupt party.

    1. tiggity Silver badge

      Re: Garbage in, garbage out

      Indeed.

      We really do not have enough of a clue about what conditions life (of some sort) can develop in - unsurprising as we only have earth life forms to study.

      e.g. lots of chemistry is "weird" at low temperatures compared to warm conditions (e.g. look at super conductor research)

      Our knowledge of low tempearture chemsitry is currenty low (had to have the bad pun!) - so it could be that chemistry quirks make life far more likely than expected in very cold conditions.. ofr I could be spouting dross, it's all guesswork and this "AI" alien life estimnate should reflect that and be treated as about as reliable as any randomly genearted number

      1. Smooth Newt
        Happy

        Re: Garbage in, garbage out

        How are they going to train the neural network? It is not as if there is a whole load of data out there on the probability of different life forms evolving in exotic atmospheres.

        Just because you write AI on the grant application doesn't mean you can perform magic.

        Talking of which, I guess grant applications will be written by AI systems in the future, and some probably are now. There's lots of data on which buzzwords etc press the right buttons. Of course it will turn into an arms race sooner with better and better research grant writing neural nets pitted against better and better research grant application assessment neural nets.

        1. DropBear Silver badge
          Trollface

          Re: Garbage in, garbage out

          Based on the laws of natural selection I predict that the first true AI will be a system developed to write convincing AI research grant applications...

          1. Flakk
            Pint

            Re: Garbage in, garbage out

            Based on the laws of natural selection I predict that the first true AI will be a system developed to write convincing AI research grant applications...

            Hilarious. And likely correct. For you.

  8. handleoclast

    Wow!

    Whoda thunk it?

    You can train an AI to make wild-arsed guesses and it will then make wild-arsed guesses that somewhat match your own wild-arsed guesses. How incredibly useful.

    1. Kyle Roberts

      Re: Wow!

      This is the most sensible answer to the question.

      Fred Hoyle, British mathematician and astronomer, said the probability of the formation of just one of the many proteins on which life depends is comparable to that of the solar system packed full of blind people randomly shuffling Rubik’s cubes all arriving at the solution at the same time.

      F. Hoyle, ‘The big bang in astronomy’, New Scientist, 92(1280):527, 1981

      GI=GO

      1. handleoclast
        Boffin

        Re: Wow!

        Fred Hoyle, British mathematician and astronomer, said the probability of the formation of just one of the many proteins on which life depends is comparable to that of the solar system packed full of blind people randomly shuffling Rubik’s cubes all arriving at the solution at the same time.

        Hoyle knew astronomy, astrophysics and mathematics well. Some people even like his SF. But here he was talking total bollocks. His entire output regarding evolution is complete and utter bollocks.

        Yes, if you're talking about nothing but random chance, then the odds are astronomical. If you're talking about evolution, they're not. And if you don't understand why they're not then you don't understand evolution. Hoyle did not understand evolution.

        Come to that, he didn't understand biochemistry, either. For example, haemoglobin isn't a unique molecule where you have to hit on that precise molecule or it's game over. Haemoglobin is different in each different species. Small differences, but it's not "everyone has to solve their Rubik's cube." Haemoglobin isn't even the best solution: in your last 7 months as an embryo you had foetal haemoglobin which is better at binding to oxygen than the adult type. Then there are the arthropods, which use copper-based haemocyanin instead of haemoglobin.

        Hoyle calculated the odds of a perfect bridge hand. One where the hands are clubs, diamonds, hearts, spades in that order. Where the first cards dealt are the aces, then the twos, then the threes, etc. The odds against that happening are enormous. Fred knew how to calculate probabilities. What he didn't mention is that every possible deal has those same odds. And what he didn't understand is that the calculation he ought to have performed were the odds of you and your partner being dealt a winning hand (50/50).

        No doubt somebody is already doing the maths. And coming up with the fact that the odds of hitting haemoglobin by nothing but random chance are mega-hyper-ultra-super-astronomical. And that even allowing for a million variants of haemoglobin/haemocyanin that still leaves the odds as hyper-ultra-super-astronomical. And that will be because they understand as little of evolution as Hoyle did.

        Single-celled animals don't have haemoglobin or haemocyanin because diffusion carries sufficient oxygen. Small multi-cellular animals don't have haemoglobin or haemocyanin either because the square-cube law means they're more "surfacey" and diffusion suffices. Water can carry around 1/20th the oxygen of blood (at human body temperature) so animals that are small enough don't need anything better than water. There's even a species of antarctic fish without haemoglobin because water dissolves more oxygen at low temperature and they are so sedentary they can get by without it.

        Up to a certain size, water is sufficient. Slightly larger than that and you need something better than water, but it doesn't need to be 20 times better (as haemoglobin is). It only needs to be fractionally better. Over time, your descendants may be slightly larger than you and need something slightly better to carry oxygen than you had. Enough time, and enough size increases and you need, and end up with, via a series of small improvements accumulating over time haemoglobin or haemocyanin. Never, at any stage, was there a jump from getting by with water to your children needing haemoglobin.

        Hoyle made a straw man. Kicked the arse out of it. Set fire to it. Stamped the ashes into the ground. Then pissed on the ground. The creationists praise him for demolishing evolution, but he didn't. He destroyed a straw man. He probably sincerely thought what he was attacking was evolution, but it wasn't. Hoyle's comments regarding evolution are unmitigated bollocks.

  9. Doctor_Wibble
    Alien

    The probability is 0 until it is 1

    Because until we know it's known, it's a known unknown, as Mr Cheney might have said.

    The best way to be certain is to broadcast our existence with a very simple 'We are ready to be enslaved/crushed/eaten' and the question will resolve itself soon enough.

    If there's such a thing as friendly aliens that don't want to do any of those things they will turn up and ask us nicely to turn off the message and hope nobody heard it. Unless they also see some of our TV broadcasts in which case they will probably throw their hands/tentacles up in despair and leave us to it.

    1. Simon Ward

      Re: The probability is 0 until it is 1

      "Because until we know it's known, it's a known unknown, as Mr Cheney might have said."

      He might have, had Rumsfeld not beaten him to it ...

      1. Doctor_Wibble
        Facepalm

        Re: The probability is 0 until it is 1

        Dammit sods feck arse bollocks, you're right!

        Worse, I already knew that so why TF I put Cheney there I don't know, though I suspect I'm not the first person ever to make that exact remark about him...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019