Towards Robust Machine Learning for Transportation Systems
Seminar | October 4 | 4 p.m. | 290 Hearst Memorial Mining Building
Justin Dauwels, Nanyeng Technological University
Nanyeng Technological University's Justin Dauwels will present Towards Robust Machine Learning for Transportation Systems on Oct. 4, 2019 at 4 p.m. in 290 Hearst Memorial Mining Building at the ITS Transportation Seminar.
Abstract: The field of machine learning has progressed rapidly in the recent years, fueled especially by new developments in deep learning. While such technologies are often hyped in the media, weaknesses of deep learning systems are starting to become obvious, potentially spelling trouble for mission-critical systems. Most current deep learning systems are brittle, since they typically do not encode or learn information about the physical world. For instance, state-of-the-art deep learning based object detection systems can potentially distinguish hundreds of animals, but do not necessarily know that birds fly or fish swim. In that sense, they are far from intelligent. The next generation of deep learning systems will be more robust, by letting them learn about the physical world. How such prior information can be encoded into the deep learning networks is an emerging area of research.
In recent work, we have shown that convolutional neural networks for objection detection in images can be made substantially more robust to image transformations (occurring in real-world applications) and to adversarial attacks by incorporating prior knowledge about the physical world. We encode physical properties of objects by means of hidden variables, and let the model infer what physical transformations have taken place in a given scene. As an illustration, we will present the Afﬁne Disentangled Generative Adversarial Network (ADIS-GAN). On the MNIST dataset, ADIS-GAN can achieve over 98 percent classiﬁcation accuracy within 30 degrees of rotation, and over 90 percent classiﬁcation accuracy against FGSM and PGD adversarial attack, outshining systems trained through data augmentation.
We will also briefly outline ongoing application-oriented machine learning projects in our team related to intelligent transportation systems. At the end of the talk, we will explore future research directions.
Bio: Dr. Justin Dauwels is an Associate Professor of the School of Electrical and Electronic Engineering at the Nanyang Technological University (NTU) in Singapore. He also serves as Deputy Director of the ST Engineering NTU corporate lab, which comprises 100+ PhD students, research staff and engineers, developing novel autonomous systems for airport operations and transportation.
His research interests are in data analytics with applications to intelligent transportation systems, autonomous systems, and analysis of human behaviour and physiology. He obtained his PhD degree in electrical engineering at the Swiss Polytechnical Institute of Technology (ETH) in Zurich in December 2005. Moreover, he was a postdoctoral fellow at the RIKEN Brain Science Institute (2006-2007) and a research scientist at the Massachusetts Institute of Technology (2008-2010). He has been a JSPS postdoctoral fellow (2007), a BAEF fellow (2008), a Henri-Benedictus Fellow of the King Baudouin Foundation (2008), and a JSPS invited fellow (2010, 2011). He is quite active in the IEEE community, as conference chair, associate editor, and other roles. He is co-founder of the spin-off companies Vigti and Mindsigns Health.
His research on intelligent transportation systems has been featured by the BBC, Straits Times, Lianhe Zaobao, Channel 5, and numerous technology websites. Besides his academic efforts, the team of Dr. Justin Dauwels also collaborates intensely with local start-ups, SMEs, and agencies, in addition to MNCs, in the field of data-driven transportation, logistics, and digital health.