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 ...
In an ever more digital world, pen and paper are increasingly getting replaced with screens and keyboards in classrooms. Now, a new study has investigated neural networks in the brain during hand- and ...
Abstract: We propose a technique for producing `visual explanations' for decisions from a large class of Convolutional Neural Network (CNN)-based models, making them more transparent. Our approach - ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback