Latest adversarial ai publications

Machine learning researchers are not solely concerned with improving the accuracy of models. They want to know how they can be corrupted and undermined – a research agenda that warrants more attention. Techerati spoke to an IBM researcher at the heart of it.