Thanks to Phil Krause for this post
What’s the fastest way to find Wally, or Waldo as he’s inexplicably called in the US and Canada? Why, with machine learning, of course. The new method derived by a doctoral student at Michigan State University’s High-Performance Computing Centre, entirely defeats the simple pleasure of aimlessly staring at a page covered in animated characters that look like Wally, but are in fact imposters planted by fiendish Martin Handford, creator of the Where’s Wally books.
“I was going to pull out every machine learning trick in my tool box to compute the optimal search strategy for finding Waldo. I was going to crush Slate’s supposed foolproof strategy and carve a trail of defeated Waldo-searchers in my wake,” Randy Olson posted on his blog.
Olson had been trapped at home by a snow storm, which gave him time to spare. He used that time to derive the method for finding Wally fastest using a data set mapping out all of Wally’s positions across all 68 of Handford’s books published since 1987. The co-ordinates where plugged into a genetic algorithm – a search process that mimics the process of natural selection used in artificial intelligence as part of machine learning. The result was a trial and error search that optimised the fastest route to take across the page to find Wally.
“Genetic algorithms continually tinker with the solution — always trying something slightly different from the current best solution and keeping the better one — until they can’t find a better solution any more,” explained Olson. He ran the algorithm for five minutes and found the solution – a path across the page that takes into account that Wally is never in the top left of the page, as he would be covered by the postcard, or in the bottom right as that’s the most looked at part of the entire page, as readers see it first when turning the pages of the book. “This path represents one of the shortest possible paths to follow on the page to find Waldo, so if we followed this path exactly, we’d most likely find Waldo much faster than someone following a more basic technique,” said Olson.
The method has yet to be put to the test, but should you really need to find Wally fast, rather than enjoy the search, this could save yours or Wally’s bacon. Via Where’s Wally? There’s an algorithm for that