Shreyank N Gowda

By explicitly aligning semantic representations across modalities, SECOS enables open-world semi-supervised learning models to directly predict textual labels without post-processing, achieving state-of-the-art classification accuracy.

Training deep nets doesn't need to be a data deluge: dynamically dropping less-useful training examples during learning can maintain accuracy while slashing compute.

Image compression, a seemingly benign process, can drastically amplify the power of adversarial attacks, making your image classifiers far more vulnerable than you thought.