Algorithms + Data Structures = Bollocks
… more often than you’d think.
I’d never actually heard of algorithms until many moons ago when, as a 23-year-old zoology graduate, I needed to bring my maths, stats and computing up to speed if I was to get the job I was after as a fisheries scientist.
So, being a glutton for punishment, I signed up for the most masochistic educational experiences of my life aka the University of York’s MSc in Biological Computation.
One of the books on its reading list was entitled ‘Data Structures + Algorithms = Programs’. Well, that equation was news to me, but despite my less-than-mathematical brain, I somehow managed to pass the course and went forth having successfully added ‘algorithm’ to my lexicon.
Scroll forward many years, and algorithms seem to rule everything from Little Britain’s ‘Computer says “No”‘ to recommendations for pages or ‘friends’ to follow on social media. A bank manager won’t refuse you a loan, but an algorithm will!
Of course algorithms can be seriously flawed. As the Graun reported nearly a decade ago:
“the very feature that makes algorithms so valuable – their ability to replicate human decision-making in a fraction of the time – can be a double-edged sword. If the observed human behaviours that dictate how an algorithm transforms input into output are flawed, we risk setting in motion a vicious circle when we hand over responsibility to The Machine”.
The most notable flaws appear to be that ‘human behaviours’ are often biased, sexist, racist or as many other ‘ist’s, subconcious or otherwise, that you can think of. That’s not something to desire in an objective decision-making system, particularly when, as last year in the UK, many students’ exam outcomes were downgraded based on their schools’ past results despite their own, individual strong performance.
On a lighter note, the weakness of algorithms can be seen in some of the recommendations made to me on social media.
As a bit of a canal anorak, I follow the ‘Foxes Afloat’ Instagram page and their YouTube channel and I have occasionally ‘liked’ one of their posts. So why, as a benign anti-monarchist, does my ‘like’ of an Instagram post about a gay couple living aboard a narrowboat with their cocker spaniel Otis, result in a recommendation for an Instagram post that shows HRH the Prince of Wales (alongside the present Mrs Wales) simply because you liked a post from foxesafloat?
Or why should Facebook feel a recommendation is in order for me to join a Facebook Group of manicure enthusiasts, or one of horse lovers, simply because some friends are members WHEN I DON’T HAVE ANY FACEBOOK FRIENDS!
(I don’t like Facebook, I’m on it, or rather my dog is, because it’s the only way to receive notifications about her agility class and that’s the way it will stay because although I have friends that are on Facebook, they are precisely that, friends, and Facebook has nothing to do with it!)
So, today’s limerick is a bit of a rant…
Insta’s ‘recommend’ algorithm
Deserves to be met with derision
‘Cos I always dislike
What it thought that I’d like.
It’s just guesswork that’s lacking precision!