In the bigera where human intelligence is being enhanced by machine intelligence through artificial intelligence (AI), machine learning (ML), big data and computer vision. However, prediction models trained using in-house training dataset suffer from two well-known problems: (1) The prediction accuracy of a trained model heavily depends on the generalization of the training dataset and may suffer from poor accuracy for unseen data inputs. (2) A privately trained prediction model is vulnerable to adversarial inputs, which manipulates the prediction outputs with only a black box access to a learning as a service API, turning a learning system against itself through input data poisoning attacks. In this talk, I will describethe formal metrics to quantitatively evaluate and measure the robustness of a trained prediction model against unseen inputs in the presence of different adversarial settings and share some of our current research results.An important takeaway message is that the defensemechanisms for guardingthe robustness of a deep learning system should be geared towards improving the generalization properties of the target learning system.
Prof. Dr. Ling Liu is a Professor in the School of Computer Science at Georgia Institute of Technology. She directs the research programs in Distributed Data Intensive Systems Lab (DiSL), examining various aspects of large-scale data intensive systems. Prof. Liu is an internationally recognized expert in the areas of Big Data Systems and Analytics, Distributed Systems, Database and Storage Systems, Internet Computing, Privacy, Security and Trust. Prof. Liu has published over 300 international journal and conference articles, and is a recipient of the best paper award from a number of top venues, including ICDCS 2003, WWW 2004, Pat Goldberg Memorial Best Paper Award 2005, IEEE CLOUD 2012, IEEE ICWS 2013, ACM/IEEE CCGrid 2015, IEEE Edge 2017, IEEE ICIOT 2017. Prof. Liu has served as general chair and PC chairs of numerous IEEE and ACM conferences in the fields of big data, cloud computing, data engineering, distributed computing, very large databases, World Wide Web, and served as the editor in chief of IEEE Transactions on Services Computing from 2013-2016. Currently Prof. Liu is co-PC chair of The Web 2019 (WWW 2019) and the Editor in Chief of ACM Transactions on Internet Technology (TOIT). Prof. Liu is an elected IEEE Fellow and a recipient of IEEE Computer Society Technical Achievement Award 2012.