Speaker: Dr. Jianjun Hu Abstract: Deep learning has led to tremendous progress in computer vision, speech recognition, and natural language processing. It has now crossed the boundary and has brought breakthroughs also in the area of bioinformatics. One interesting problem is developing accurate models for predicting peptide binding affinities to protein receptors such as MHC(Major Histocompatibility complex), which can shed understanding to adverse drug reaction and autoimmune diseases and lead to more effective protein therapy and design of vaccines. We proposed a deep convolutional neural network (CNN) based peptide binding prediction algorithm for achieving substantially higher accuracy as tested in MHC-I peptide binding affinity prediction. Our model takes raw binding peptide sequences and affinity scores or binding labels as input without needing any human-designed features. The back-propagation training algorithm allows it to learn nonlinear relationships among the amino acid positions of the peptides. It also can naturally handle the peptide length variation, MHC polymorphasim, and unbalanced training samples of MHC proteins with different alleles via a simple amino acid padding scheme. Our experiments showed that DeepMHC can achieve the state-of-the-art prediction performance on most of the IEDB benchmark datasets with a single model architecture and without using any consensus or composite ensemble classifier models. Bio: I joined CSE department of the University of South Carolina in August 2007. I am now working on integrative functional genomics and especially integrative analysis of microarray data. I am also interested in motif discovery for understanding gene expression mechanisms involved in diseases. I got my Ph.D. in Computer Science in the area of machine learning and particularly evolutionary computation at the Genetic Algorithm Research and Application Group (GARAGe) of Michigan State University. My dissertation focuses on sustainable evolutionary computation algorithms and automated computational synthesis. I have worked on the DNA motif discovery problem as Postdoc at Kihara Bioinformatics Lab, Purdue University and microarray analysis at the Computational Molecular Biology Division at the University of Southern California (another USC).

Abstract: Patching is one of the most important measures to continuously uphold security throughout the life of a software system. Patch generation and deployment are probably the most critical tasks in the process of patching. However, patch generation is typically a lengthy procedure (according to Symantec, it takes an average of 28 days to release a patch for fixing a critical security bug); and patch deployment risks system stability due to new bugs contained in patches. From the perspective of speeding up patch generation and avoiding bugs in patches, we examine the notorious heap vulnerabilities, including heap buffer overflows (such as Heartbleed), uninitialized-read, use-after-free, and double-free, and explore the following two important but less-investigated problems: (1) How fast can heap patches be generated? (2) How to ensure zero bugs in the generated patches? While quick patch generation and bug-free patches are two naturally desired goals, they contradict each other in practice. Rushed patch generation tends to introduce bugs, while creating a quality patch requires significant time for debugging, testing, and even system redesign. Thus, how to achieve the two inherently contradictory goals simultaneously has been challenging. Inspired by “targeted therapy”, a cancer treatment that precisely recognizes and kills cancer cells, we propose Targeted Heap Therapy to pinpoint and treat vulnerable buffers, which are buffers that can be exploited to launch attacks, with instantly generated bug-free patches. This talk will also present some of the important problems and future prospects on Internet of Things security as well as our ongoing work on these problems. Bio: Dr. Qiang Zeng is a Tenure-Track Assistant Professor in the Department of Computer & Information Sciences at Temple University. He received his Ph.D. in Computer Science and Engineering from the Pennsylvania State University. He has rich industry experience and has worked in the IBM T.J. Watson Research Center, the NEC Lab America, Symantec and Yahoo. Dr. Zeng’s main research interest is Systems and Software Security. He currently works on Mobile Security, IoT Security, and deep learning for solving security problems. He has published papers in PLDI, NDSS, MobiSys, CGO, DSN and TKDE.

https://www.facebook.com/events/386161698480048/

THESIS DEFENSE Department of Computer Science and Engineering University of South Carolina Author : Meng Ying Advisor : Dr. Gregory Gay Abstract Test adequacy criteria are widely used to guide test creation. However, many of these criteria are sensitive to statement structure or the choice of test oracle. This is because such criteria ensure that execution reaches the element of interest, but impose no constraints on the execution path after this point. We are not guaranteed to observe a failure just because a fault is triggered. To address this issue, we have proposed the concept of observability—an extension to coverage criteria based on Boolean expressions that combines the obligations of a host criterion with an additional path condition that increases the likelihood that a fault encountered will propagate to a monitored variable. Our study, conducted over five industrial systems and an additional forty open-source systems, has revealed that adding observability tends to improve efficacy over satisfaction of the traditional criteria, with average improvements of up to 392.44% in mutation detection and per-model improvements of up to 1654.38%. Ultimately, there is merit to our hypothesis—observability reduces sensitivity to the choice of oracle and to the program structure.

I would like to invite you to attend this week's CSCE 791 seminar. These seminars highlight research being performed in our department and across the world. All CSCE 791 seminars are open to anybody who wishes to attend - not just students registered for the course. Friday, November 10, 2:20 - 3:10 PM Swearingen room 2A14 Speaker: Nikolaos Vitzilaios, University of South Carolina (Department of Mechanical Engineering) Abstract: The area of Unmanned Systems & Robotics has seen a tremendous growth over the last decades with autonomous systems being rapidly developed in many domains resulting in a wide range of applications that we can see in our daily lives (drones, autonomous cars, industrial robots, medical and service robotics, space robotics, etc.). This presentation will show the research of Dr. Nikolaos (Nikos) Vitzilaios in this area over the last 10 years, presenting the developments in specific areas over time and focusing on latest research as long as future research plans. The presentation will include research in the following areas: Aerial Robotics: applications of automatic control in unmanned fixed-wing aircraft and rotorcraft, including theoretical aspects as well as applied hardware and software developments. Several platforms have been developed over the last 10 years while the latest development will be presented based on a patented design of a dual-tilting quadcopter able to perform advanced navigation and control in narrow spaces as well as fault-tolerant control. Mobile Robotics: several ground robotic platforms will be presented, including customized commercial mobile robots as well as in-house built robots (including a patented one). These platforms are built for different applications and projects and the presentation will focus on the collaboration with aerial robots in critical missions UAV Aerodynamics: novel research in the area of circulation control wings for fixed-wing aircraft will be presented. Marine Robotics: a novel propulsion system will be presented for low speed propeller less robots that are required to be used in extreme environments (nuclear reactors). Medical Robotics: the latest research on the modelling of the human thumb will be presented accompanied by a new kinematic model that shows the importance of the thumb in grasping and how this will affect our perception for the development of future robotic hands. Mechatronic Systems: the development of an automatic bike gear shifter (patent pending). Modeling and control of complex and highly nonlinear systems. Future trends and planned research in perception and control. The presentation will focus on the outputs of each research project and will include demos and videos from field experiments and indoor-outdoor testing. Bio: Dr. Nikolaos (Nikos) Vitzilaios is an Assistant Professor at the Department of Mechanical Engineering, University of South Carolina, since August 2017. He holds a PhD in Mechanical Engineering from the Technical University of Crete (2010) and his PhD thesis was on the development of autonomous controllers for helicopter UAVs. Prior to joining USC, he was a Senior Lecturer in Robotics at Kingston University, London, UK. From 2011-2012, he was a Postdoctoral Fellow at the Department of Electrical Engineering, University of Alberta, Canada, working in the Applied Nonlinear Controls Laboratory and developing a helicopter UAV for power line inspection, funded by the Canadian government (NSERC). From 2012-2015, he was a Research Scientist at the University of Denver Unmanned Systems Research Institute (Department of Electrical Engineering), leading the Aerial Robotics Team and working on several projects in the area of unmanned systems funded by various agencies (NSF, ARL, NASA). Dr. Vitzilaios has more than 10 years of research and more than 5 years of teaching experience in the areas of Robotics and Controls, with notable presence in the Robotics & Automation society, more than 30 publications, one US patent and successful grant applications both in US and UK. His background is interdisciplinary from the areas of Mechanical Engineering, Electrical Engineering and Computer Science. His research interests span the broad area of Autonomous Unmanned Systems where he has significant hands-on experience in all kinds of robotic applications (aerial, ground, marine, industrial, biomedical). His research is mainly experimental and his interests include prototype development and commercialization of research outcomes. He is a Fellow of the Higher Education Academy in UK and a member of IEEE, AIAA, AUVSI and IFAC. He is a Chartered Mechanical Engineer in the Technical Chamber of Greece since March 2005.

THESIS DEFENSE Department of Computer Science and Engineering University of South Carolina Author : Adel D. Rajab Advisor : Dr. Chin-Tser Huang Date : Nov 10th 2017 Time : 2:00 pm Place : Meeting room 2265, Innovation Center Abstract The Optical Bust Switching (OBS) network has become one of the most promising switching technologies for building the next-generation Internet backbone infrastructure. However, OBS networks still face a number of security and Quality of Service (QoS) challenges, particularly from Burst Header Packet (BHP) flooding attacks. First, we proposed and developed a new security model that can be embedded into OBS core switch architecture to prevent BHP flooding attacks. The countermeasure security model allows the OBS core switch to classify the ingress nodes based on their behavior and the amount of reserved resources not being utilized. A malicious node causing a BHP flooding attack will be blocked by the developed model until the risk disappears or the malicious node redeems itself. Using our security model, we can effectively and preemptively prevent a BHP flooding attack regardless of the strength of the attacker. In the second part of this dissertation, we investigated the potential use of machine learning (ML) in countering the risk of the BHP flood attack problem. In particular, we proposed and developed a new series of rules, using the decision tree method to prevent the risk of a BHP flooding attack. The proposed classification rule models were evaluated using different metrics to measure the overall performance of this approach. The experiments showed that using rules derived from the decision trees did indeed counter BHP flooding attacks, and enabled the automatic classification of edge nodes at an early stage. In the third part of this dissertation, we performed a comparative study, evaluating a number of ML techniques in classifying edge nodes, to determine the most suitable ML method to prevent this type of attack. The experimental results from a preprocessed dataset related to BHP flooding attacks showed that rule-based classifiers, in particular decision trees (C4.5), Bagging, and RIDOR, consistently derive classifiers that are more predictive, compared to alternate ML algorithms, including AdaBoost, Logistic Regression, Naive Bayes, SVM-SMO and ANN-MultilayerPerceptron. Moreover, the harmonic mean, recall and precision results of the rule-based and tree classifiers were more competitive than those of the remaining ML algorithms. Lastly, the runtime results in ms showed that decision tree classifiers are not only more predictive, but are also more efficient than other algorithms. Thus, our findings show that decision tree identifier is the most appropriate technique for classifying ingress nodes to combat the BHP flooding attack problem.

THESIS DEFENSE Department of Computer Science and Engineering University of South Carolina Author : Tian Jing Advisor : Dr. Wenyuan Xu Date : Nov 6th 2017 Time : 12:00 pm Place : Meeting room 2265, Innovation Center. Abstract Authentication plays a key role in securing various resources including corporate facilities or electronic assets. As the most used authentication scheme, knowledge-based authentication is easy to use but its security is bounded by how much a user can remember. Biometrics-based authentication requires no memorization but ‘resetting’ a biometric password may not always be possible. Thus, we propose study several behavioral biometrics (i.e., mid-air gestures) for authentication which does not have the same privacy or availability concerns as of physiological biometrics. In this dissertation, we first propose a user-friendly authentication system KinWrite that allows users to choose arbitrary, short and easy-to-memorize passwords while providing resilience to password cracking and password theft. Specifically, we let users write their passwords (i.e., signatures in the 3D space), and verify a user’s identity with similarities between the user’s password and enrolled password templates. Dynamic time warping distance is used for similarity calculation between 3D passwords samples. In the second part of the dissertation, we design an authentication scheme that does not depend on the handwriting contents, i.e., regardless of the written words or symbols, and adapt challenge-response mechanism to avoid possible eavesdropping, man-in-the-middle attacks, and reply attacks. We design a MoCRA system that utilizes Leap Motion to capture users’ writing movements and use writing style to verify users, even if what they write during the verification is completely different from what they write during the enrollment. Specifically, MoCRA leverages co-occurrence matrices to model the handwriting styles, and use a Support Vector Machine (SVM) to accept a legitimate user and reject the rest. In the third part, we study both security and usability performance on multiple types of mid-air gestures that used as passwords, including writing signatures in the air. We objectively quantify the usability performance by metrics related to the enroll time and the complexity of the gestures, and evaluate the security performance by the authentication performance. In addition, we subjectively evaluate the gestures by survey responses from both field subjects who participated in gesture experiments and on-line subjects who watched a short video on gesture introducing. Finally, we study the consistency of gestures over samples collected in a two-month period, and evaluate their security under shoulder surfing attacks.

I would like to invite you to attend this week's CSCE 791 seminar. These seminars highlight research being performed in our department and across the world. All CSCE 791 seminars are open to anybody who wishes to attend - not just students registered for the course. Speaker: Bryant Walker Smith, University of South Carolina (School of Law) Abstract: This discussion will explore the technologies, applications, and legal aspects of automated driving. Bio: Bryant Walker Smith is an assistant professor in the School of Law and (by courtesy) in the School of Engineering at the University of South Carolina. He is also an affiliate scholar at the Center for Internet and Society at Stanford Law School, chair of the Emerging Technology Law Committee of the Transportation Research Board of the National Academies, and a member of the New York Bar. Bryant's research focuses on risk (particularly tort law and product liability), technology (automation and connectivity), and mobility (safety and regulation). As an internationally recognized expert on the law of self-driving vehicles, Bryant taught the first-ever course on this topic and is regularly consulted by government, industry, and media. His recent article, Proximity-Driven Liability, argues that commercial sellers' growing information about, access to, and control over their products, product users, and product uses could significantly expand their point-of-sale and post-sale obligations toward people endangered by those products. Before joining the University of South Carolina, Bryant led the legal aspects of automated driving program at Stanford University, clerked for the Hon. Evan J. Wallach at the United States Court of International Trade, and worked as a fellow at the European Bank for Reconstruction and Development. He holds both an LL.M. in International Legal Studies and a J.D. (cum laude) from New York University School of Law and a B.S. in civil engineering from the University of Wisconsin. Prior to his legal career, Bryant worked as a transportation engineer.

Duncan Buell The talk will focus on two main topics:

• The security of elections in a networked world, where complicated computer hardware is used in the stressful circumstances of a major election, and the ability (or inability) of verifying that voter registration databases and election results have not been corrupted by malware or hacking.
• The post-election evidence based analysis and auditing of the results to provide increased confidence that the certified election results are in fact correct.

Bio: Dr. Buell is the NCR Professor of Computer Science and Engineering at USC and has been working for more than ten years on issues of election integrity, computer security for elections, and the post-election analysis of the data from ES&S iVotronic voting computers (used in South Carolina and several other states). He has since 2010 had programs to analyze the South Carolina election data to verify correctness of the results and adherence to the official protocol for conducting elections and returning the data to county headquarters for tallying. He has testified as an expert witness in open records cases and been involved in election integrity litigation in Arizona, Colorado, Georgia, North Carolina, Pennsylvania, and Texas.