Abstract: We present Momentum Contrast (MoCo) for unsupervised visual representation learning. From a perspective on contrastive learning as dictionary look-up, we build a dynamic dictionary with a ...

There’s a moment with most foldables where the magic trick stops working. It’s usually day three or four, right after the honeymoon phase, when the device stops being “the future” and starts being “a ...

Abstract: The class of random features is one of the most popular techniques to speed up kernel methods in large-scale problems. Related works have been recognized by the NeurIPS Test-of-Time award in ...