[Photo: Ryan Anderson]
Sorry dads everywhere, but Ryan Anderson has officially outdone you.
The solution architect for IBM Watson used the program to build a real life “Harry Potter” sorting hat for his two daughters, Lucy, 8, and Julia, 6.
The hat works just like in the books – you place it on your head (although this isn’t necessary) and speak to it, telling the sorting hat a a few personal things about who you are.
The sorting hat will then determine whether you’re a brave Gryffindor, wise Ravenclaw, loyal Hufflepuff or an ambitious Slytherin by using Watson’s Natural Language Classifier.
The program associates certain words it picks up to specific houses in a process called “setting a ground truth.” It’s a simple two-column system with a list of adjectives in one column and a corresponding house in the other. The program is so simple, Anderson even enlisted the help of his daughter Lucy, who created 150 “ground truths.”
The hat also works using information gleaned from a person’s Wikipedia page, sorting Hilary Clinton and Stephen Hawking into Ravenclaw with 92 and 97 per cent certainty, respectively. Disconcertingly, it places Donald Trump in Gryffindor, although with just 48 per cent certainty.
That said, the hat uses deep learning, which is an artificial intelligence that allows machines to “learn” how to complete tasks independently. As the sorting hat scans the web, more ground truths will be added resulting in more accurate predictions.
What house do you think you’d be sorted into? Let us know by tweeting to us @YahooStyleCA.