• Question: How would you make the computers know how to stop the bridge fall down ?

    Asked by zest360ear to Valerio, Sreejita, Sam, Kate, Anastasia, Adam on 9 Nov 2019.
    • Photo: Sreejita Ghosh

      Sreejita Ghosh answered on 9 Nov 2019: last edited 9 Nov 2019 12:31 pm

      I am also interested in knowing the answer to this question, hence commenting 🙂

    • Photo: Kate Winfield

      Kate Winfield answered on 11 Nov 2019:

      I have no idea 🙂 I would of thought you could computer model a bridge!

    • Photo: Anastasia Aliferi

      Anastasia Aliferi answered on 11 Nov 2019:

      I think you would give the computer examples of bridges that fell down and examples of bridges that did not and let it work out the pattern behind what makes a bridge stay up.

    • Photo: Adam Wootton

      Adam Wootton answered on 11 Nov 2019:

      I use something called a neural network, which is a bit like a computer version of a human brain. They are nowhere near as good as your human brain, which knows how to do lots of different things, such as breathing, walking and doing maths (sometimes all at the same time). Neural networks normally only know how to do one thing, but are very good at that one thing. They’re really, really good at spotting patterns.
      When you’re in school, your teachers give you lots of example questions to help you learn. You actually learn how to do everything by just trying it out and practicing. Neural networks are exactly the same. If I want to make a network to stop a bridge falling down, I will need it to know what a bridge looks like when it’s starting to get damaged. To do this, I will show it lots and lots of examples of the bridge when it is normal and when it is damaged. This teaches it to tell the difference between a working bridge and a damaged bridge.
      The computer will then get lots of information from the bridge every day and if it sees that it starts to look damaged, then it can tell somebody who needs to know about it and they can make sure that it gets repaired.