Kaggle, an online data science community that regularly hosts machine learning competitions with prizes often in the tens of thousands of dollars, has uncovered a cheating scandal involving a winning team. The Google subsidiary announced that the winner of a competition involving a pet adoption site had been disqualified from the contest for fraudulently obtaining and obscuring test set data.
The competition asked contestants to develop algorithms to predict the rate of pet adoption based on pet listings from PetFinder.my, a Malaysian pet adoption site. The goal was to help discover what makes a shelter pet's online profile appealing for adopters.
The algorithm from BestPetting, the first place team, seemed to almost perfectly predict the rate of adoption for the test set against which the submissions were evaluated. The team of three was awarded $10,000.
Nine months later a teenager found that the BestPetting team a had obtained PetFinder.my's testing data then encoded and decoded that data into their algorithm to obfuscate their advantage.
"It is very sad indeed that such brilliant people . . . have gone to such lengths to defraud a welfare competition aimed at saving precious animal lives, solely for their own financial gain," said PetFinder.my founder Andy Koh.
View Full Article
No entries found