The CyLab Distinguished Speaker Seminar series brings world-class academics, entrepreneurs, government officials, and technologists across a variety of security and privacy disciplines to engage with the CMU community.
Doors open at 11:30, and lunch will be served.
Speaker: Conrad Tucker, Professor, Mechanical Engineering, Carnegie Mellon University
Title: From Generative Neural Networks to Social Media Networks: Ascertaining the Veracity and Security of Data in the Information Age
Ascertaining the veracity and security of data in the information age is a challenge both for humans (e.g., communicating within social media networks) and machines (e.g., training data for artificial neural networks). A lack of data veracity has the potential to “fool” both machines, as well as humans into achieving different outcomes/output. From a machine learning perspective, “fooling” a machine has had a positive impact in the development of algorithms such as generative adversarial networks (GANs), and has resulted in the ability of machines to generate hyper-realistic data such as images and text. However, adverse effects can be observed in large-scale social media networks, where the veracity of data cannot be quickly ascertained. Misinformation that is spread via social media networks can result in echo-chambers, lone communities that facilitate selective content diffusion as a result of user polarization. Ironically, this misinformation can now be reliably generated using machine learning algorithms such as GANs. In this work, a network representation learning algorithm called Deep Spectral Learning is presented that characterizes echo chambers. The Deep Spectral Learning algorithm performs data pre-processing on the basis of underlying social dynamics, while learning to classify users using a deep learning classifier. The proposed Deep Spectral Learning method is able to discover echo-chambers that are formed within social media networks, hereby potentially mitigating the data veracity and security challenges. This research explores the future of human-machine learning and the challenges and opportunities that exist in information creation and exploitation.
Conrad Tucker is an Arthur Hamerschlag Career Development Professor of Mechanical Engineering and Machine Learning (courtesy) at Carnegie Mellon University. His research focuses on the design and optimization of systems through the acquisition, integration and mining of large scale, disparate data.
Tucker has served as PI/Co-PI on federally/non-federally funded grants from the National Science Foundation (NSF), the Air Force Office of Scientific Research (AFOSR), the Defense Advanced Research Projects Agency (DARPA), the Army Research Laboratory (ARL), the Office of Naval Research (ONR) via the NSF Center for eDesign, and most recently, the Bill and Melinda Gates Foundation (BMGF). He is currently serving as Co-PI in the ARL Cyber CRA at CMU. In February 2016, he was invited by National Academy of Engineering (NAE) President Dan Mote to serve as a member of the Advisory Committee for the NAE Frontiers of Engineering Education (FOEE) Symposium. He received his Ph.D., M.S. (industrial engineering), and MBA degrees from the University of Illinois at Urbana-Champaign, and his B.S. in Mechanical Engineering from Rose-Hulman Institute of Technology.