Question: How Do You Make Computers Think For Themselfs?

Andrew Hearn answered on 20 Mar 2012:

Hi there again! I’m glad that you’ve asked here, great! As it is a question that I can’t give a short answer to!

So are you sitting comfortably? I’ll begin….. 🙂

There are loads of different ways of getting computers to think for themselves, the subject of artificial intelligence is still so new – there are loads of people in universities all over the world still trying out different ways of making machines think more like we do, getting a bit nearer to sci-fi(!) – but some other people think this is scary, fair enough, quite understandable!

(What do you think of it generally? Are making robots more intelligent a good or bad thing? My dad isn’t so sure, my mum thinks the idea is scary, but nearly all of us who are involved with software think it is a great idea!)

Here at my work, we have loads of different algorithms to deal with whatever input we get (it could be the position of a machine on the ground, or radar data, or anything really), and we get processors to change what they are doing depending on these inputs and the results of the algorithms we use (“neural networking” is an example).

The ‘clever’ bit is to get the results of what the processors have calculated (plus adjusting the position of machine or other data or whatever) to become the new inputs, it goes on and on in a feedback loop. This is what we call “thinking” (as nearly all neuroscientists and programmers believe that is what the human brain is actually doing – forever dealing with loads and loads of feedback loops from nerve impulses to move muscles and releasing chemicals like serotonin, and (my fave!) adrenaline).

A.I. is a fascinating subject with loads of different theories, and the exciting bit is that it is all still so new, and very ‘cutting edge’ and we’re living right in it! 🙂