M.I.A. files her nails like a boss.
Early attempts at explaining this phenomenon focused on nonlinearity and overfitting. We argue instead that the primary cause of neural networks’ vulnerability to adversarial perturbation is their linear nature. Moreover, this view yields a simple and fast method of generating adversarial examples.
Stanford researchers solve the mystery of the dancing droplets. This is beautiful on multiple levels.
The Beast of Turin awakens: Fiat S76 driven for the first time in a century—and it’s very, very angry. They just don’t make ‘em like they used to. Each one of the four cylinders in this engine has more displacement than all eight cylinders in my truck put together.