Seminar on January 17, 2022

Title

Towards Self-Aware Artificial Intelligence – Lessons Learned from Optimal Estimation Theory

Speaker

Prof. Nidhal Bouaynaya, Rowan University, USA

When and Where

January 17, 2022, 16:00-17:00
Online on Zoom, see TAS Resilience Node for registration or send an email to Valeria Cardellini

Abstract

Deep neural networks (DNNs) have surpassed human-level accuracy in various fields, holding the promise of emerging technologies, such as self-driving cars and autonomous unmanned aircraft systems, smart cities infrastructure, personalized treatment in medicine, and cybersecurity. However, unlike Humans who have a natural cognitive intuition for probabilities, DNN systems - being inherently deterministic - are unable to evaluate their confidence in the decisions. To truly deserve its name, an artificial intelligence system must be aware of its limitations and have the capacity for insightful introspection.

This talk will advance Bayesian deep learning methods that are able to quantify their uncertainty in the decision and self-assess their performance, are robust to adversarial attacks, and can even expose an attack from ambient noise. This talk will establish the theoretical and algorithmic foundations of uncertainty or belief propagation through complex deep learning models by adopting powerful frameworks from optimal estimation problems in non-linear and non-Gaussian dynamical systems.

The challenge in DNNs is the multi-layer stages of non-linearities in deep learning models, which makes propagation of high-dimensional distributions mathematically intractable. Drawing upon powerful statistical frameworks for density propagation in non-linear and non-Gaussian dynamical systems, we introduce Tensor Normal distributions as priors over the network parameters and derive a first-order Taylor series mean-covariance propagation framework. We subsequently extend this first-order approximation to an unscented framework that propagates sigma points through the model layers. The unscented framework is shown to be accurate to at least the second-order approximation of the posterior distribution. We finally learn the entire predictive distribution using Particle Filtering, a powerful class of numerical methods for the solution of optimal estimation problems in non-linear, non-Gaussian systems. The uncertainty in the output decision is given by the propagated covariance of the predictive distribution. Furthermore, we show that the proposed framework performs an automatic logit squeezing, which leads to significantly enhanced robustness against noise and adversarial attacks. Experimental results on benchmark datasets, including MNIST, CIFAR-10, real-world synthetic aperture radar (SAR), and Brain tumor segmentation (BraTS 2015), demonstrate: 1) superior robustness against Gaussian noise and adversarial attacks; 2) self-assessment through predictive confidence that monotonically decreases with increasing levels of ambient noise or attack; and 3) an ability to detect a targeted attack from ambient noise.

Short Bio

Dr. Nidhal Carla Bouaynaya holds a Ph.D. in Electrical and Computer Engineering (ECE) and an M.S. in Pure Mathematics from the University of Illinois at Chicago. She is a Professor of ECE and the Director of Rowan’s Artificial Intelligence Lab (RAIL). She is currently serving as the Associate Dean for Research and Graduate Studies of the Henry M. Rowan College of Engineering. Previously, she was a faculty member with the University of Arkansas at Little Rock.

Her research interests are in Big Data Analytics, Machine Learning, Artificial Intelligence and Mathematical Optimization. She co-authored more than 100 refereed journal articles, book chapters and conference proceedings, such as IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Signal Processing Letters, IEEE Signal Processing Magazine and PLOS Medicine. Dr. Bouaynaya won numerous Best Paper Awards, the most recent was at the 2019 IEEE International Workshop on Machine Learning for Signal Processing. She is also the winner of the Top algorithm at the 2016 Multinomial Brain Tumor Segmentation Challenge (BRATS). Dr. Bouaynaya has been honored with numerous research and teaching awards, including Rowan Research Achievement Award in 2017 and The University of Arkansas at Little Rock Faculty Excellence Award in Research.

Her research is primarily funded by the National Science Foundation (NSF CCF, NSF ACI, NSF DUE, NSF I-Corps, NSF ECCS, NSF OAC and NSF HRD), The National Institutes of Health (NIH), US. Department of Education (USED), New Jersey Department of Transportation (NJ DoT), US. Department of Agriculture (USDA), the Federal Aviation Administration (FAA), Lockheed Martin Inc. and other industry. She is also interested in entrepreneurial endeavors. In 2017, she Co-founded and is Chief Executive Officer (CEO) of MRIMATH, LLC, a start-up company that uses artificial intelligence to improve patient oncology outcome and treatment response. MRIMath is funded by the NIH SBIR Program.