Schedule: Reliability and Safety sessions

Broader AI adoption and gaining trust from customers requires AI systems to be fair, interpretable, robust, and safe. Anand Rao synthesizes the current research in FAT (fairness, accountability, and transparency) into a step-by-step methodology to address these issues—illustrated with case studies from the financial services and healthcare industries. Read more.

Drawing on his work building and deploying an RL-based relational query optimizer, a core component of almost every database system, Sanjay Krishnan highlights some of the underappreciated challenges to implementing deep reinforcement learning. Read more.

Pradip Bose details a next-generation AI research project focused on creating "self-aware" AI systems that have built-in autonomic detection and mitigation facilities to avoid faulty or undesirable behavior in the field—in particular, cognitive bias and inaccurate decisions that are perceived as being unethical. Read more.

Artificial intelligence has shown great potential to revolutionize clinical medicine and healthcare delivery. However, incorporating these algorithms into clinical workflows involves a big challenge: convincing clinicians and regulators to trust a “black box” solution. Behrooz Hashemian explains how he's helping make deep neural networks interpretable to provide evidence for clinical decisions. Read more.

While deep learning has led to many advancements in computer vision, most research has focused on ground-based imagery. Ryan Mukherjee and Neil Fendley offer an overview of functional Map of the World (fMoW), an ImageNet for satellite imagery built to address this issue, and explain how you can attack or defend these deep learning models. Read more.