MIT researchers say the influential ImageNet data set has "systematic annotation issues" when used as a benchmark for evaluating object recognition models.

The tool makes small changes to images that are imperceptible to the human eye but nevertheless hamper AI-generated alterations.

Researchers from MIT have 3D printed a turtle that can fool image recognition algorithms into thinking it's a rifle. Here's why that's scary.

When something does zero-shot image classification, that means it’s able to make judgments about the contents of an image without the user needing to train the system beforehand on what to look ...