Abstract: CNNs, specialized neural networks for Computer Vision tasks, are used in sensitive contexts and exposed in the wild. While extremely accurate, they are also sensitive to imperceptible perturbations that can’t be detected by human eyes. For this reason, they have been targeted by hackers which implemented AI-based techniques for their malicious purposes. During this workshop we are going to learn some synthetic attacking techniques and a defence strategy to mitigate the effect of such attacks and make neural networks more robust to them, while at the same time keeping minimal impact on the accuracy of the model and implementation costs. We would also try to understand if Transformers applied to Computer Vision tasks are immune to Adversarial Attacks.
Bio: Guglielmo is part of MSD (Merck & Co. in North America). He is currently busy unlocking business value through Computer Vision and other ML/DL/AI applications to the biotech manufacturing space. He has an extensive background in Software Engineering and Data Science across other big organizations including IBM, Optum and FAO of the UN in diverse contexts (such as Healthcare, DevOps, Cyber Security). Guglielmo has been recognized as DataOps Champion at the Streamsets DataOps Summit 2019 and awarded as one of the Top 50 Tech Visionaries at the 2019 Dubai Intercon Conference. Since 2018 he is also an international speaker (almost 30 international conferences so far, including Big Things 2019 and 2020, Spark+AI Summit 2019, Annual Cyber Security and AI Summit 2019 and 2020), author of a tech book on distributed Deep Learning with Apache Spark and planning about a second tech book which should be probably released at the end of 2021.