University of South Carolina

Department of Computer Science & Engineering

Spring 2015             CSCE 791: Seminar in Advances in Computing

Instructor: Dr. Yan Tong,

Email: tongy@cec.sc.edu

Phone: 777- 0801

Office: Swearingen 3A52,

*Course Time and Location: Friday 2:50 pm - 4:05 pm, SWGN 2A05

*Course time and location might be changed according to announcements of Computer Science and Engineering Colloquium.

Class Homepage: http://www.cse.sc.edu/~tongy/csce791/csce791.html

Expectations and Requirements:

· This is a graduate-level seminar course. You are expected to attend and engage yourself in every talk.

· A brief written summary is required after each seminar. You should summarize the content of the talk, describe the major contributions and/or take-home messages, and make positive/negative comments on the talk. The summary is about one or two paragraphs and should not exceed one page.

· The written summary should be submitted via department dropbox before 11:59:59pm, the following Thursday.

Academic Integrity: Written summaries are expected to be the sole effort of the student submitting the work. Students are expected to follow the Code of Student Academic Responsibility. Every instance of a suspected violation will be reported. Students found guilty of violations of the Code will receive the grade of F for the course in addition to whatever disciplinary sanctions are applied.

Scheduled Presentations

Date and Location	Speaker	Topic/Abstract
2:50 pm - 4:05 pm Friday, January 16 Swearingen 2A05	Dr. Yan Tong Department of Computer Science & Engineering University of South Carolina	Title: Improving Feature Learning, Feature Selection, and Classification in Facial Expression Analysis Abstract: Facial activity is the most direct signal for perceiving emotional states in people. Emotion analysis from facial displays has been attracted an increasing attention because of its wide applications from human-centered computing to neuropsychiatry. Most recently, unsupervised, semi-supervised, and supervised feature learning approaches have been employed in capturing underlying patterns in facial images and have shown promise in facial expression recognition. In this talk, I will present one of our recent work for improving feature learning, feature selection, and classification in facial expression analysis. Extensive experiments on public databases showed that our approach yielded dramatic improvements in facial expression analysis.
9:30 am – 11:00 am Friday, January 23 Swearingen 2A11	Yingjie Lao University of Minnesota	Title: Design of Secure and Anti-Counterfeit Integrated Circuits  Abstract: As electronic devices become increasingly interconnected and pervasive in people's lives, security, trustworthy computing, and privacy protection have notably emerged as important challenges for the next decade. The assumption that hardware is trustworthy and that security effort should only be focused on networks and software is no longer valid given globalization of integrated circuits (ICs) and systems design and fabrication. The design of secure hardware ICs requires novel approaches for authentication that are based on multiple factors that is difficult to compromise. Equally important is the need for protecting intellectual property and design of integrated circuits that are harder to reverse engineer. In this talk, I will explore both of the authentication-based and obfuscation-based hardware protection approaches. One popular technique of authentication-based protection is Physical Unclonable Functions (PUFs) which provide a hardware specific unique signature or identification. I will provide an overview of the reconfigurable PUF structures and circuits that could achieve higher security. Next, I will present a novel low-overhead solution to design Digital Signal Processing (DSP) circuits that are obfuscated both structurally and functionally by utilizing high-level transformation techniques. Finally, I will describe the theme of my future research: security, randomness, and computing.
9:30 am – 10:45 am Friday, January 30 Swearingen 1A03 (Faculty Lounge)	Daniel Wong University of Southern California	Title: Energy Proportional Datacenters Abstract: Datacenters provide the infrastructure backbone necessary to support big data analytics and cloud services, which are increasingly employed to tackle a diverse set of grand challenges. But datacenter power consumption is growing at an unsustainable pace. In order to keep up with the hyperscale growth in datacenter demand, it is imperative that datacenters become more energy efficient. Servers, the largest power consumer in datacenters, are optimized for high energy efficiency only at peak and idle load, but rarely operate in that region. Therefore, there is a need for energy proportional computing, where servers consume power in proportion to their utilization. How to achieve or surpass ideal energy proportionality is the focus of this talk. Toward this goal, I will first present a historical trend analysis of energy proportionality, using novel metrics, in order to identify opportunities for proportionality improvements. Second, I will present KnightShift, a heterogeneous server architecture that tightly couples a low-power Knight node with a high-power primary server, which achieves near-ideal energy proportionality. Finally, I will present the implications of high energy proportional servers on cluster-level energy proportionality. We find that traditional cluster-level energy proportionality techniques may actually limit cluster-wide energy proportionality, and it may now be more beneficial to depend solely on server-level low power techniques such as KnightShift; a finding that is a major departure from conventional wisdom.
9:30 am – 10:45 am Friday, February 6 Swearingen 1A03 (Faculty Lounge)	Mai Zheng The Ohio State University	Title: Torturing Storage Systems for Fun and Profit Abstract: Storage system failures are extremely damaging — if your browser crashes you sigh, but when your family photos disappear you cry. So we need highly reliable storage systems that can keep data safe even under failures. Such high standard of reliability is far from trivial to provide, particularly when high performance must be achieved. This leads to complex and error-prone code—even at a low defect rate of one bug per thousand lines, the millions of lines of code in a commercial online transaction processing (OLTP) database can harbor thousands of bugs.  In this talk, I will focus on two of my research efforts to better understand the reliability of data storage systems under failures. First, I will discuss a framework for evaluating solid-state drives (SSDs). This work uncovers five failure behaviors of SSDs, including bit corruption, shorn writes, non-serializable writes, metadata corruption, and total device failure. The surprising results provide important implications to the design of higher-level storage software and have led to the enhancement of power loss protection in some latest SSDs. In the second part, I will detail a framework to expose and diagnose atomicity, consistency, isolation, and durability (ACID) violations in databases under failures. Using the framework, we study eight widely-used databases. Surprisingly, all eight databases exhibit erroneous behavior. For the open-source databases, we are able to diagnose the root causes using our framework, and for the proprietary commercial databases we can reproducibly induce data loss.
9:30 am – 10:45 am Friday, February 13 Swearingen 1A03 (Faculty Lounge)	Greg Gay University of Minnesota	Title: Steering Model-Based Test Oracles to Admit Real Program Behaviors Abstract: There are two key artifacts necessary to test software, the test data - inputs given to the system under test (SUT) - and the oracle - which judges the correctness of the resulting execution. Substantial research efforts have been devoted towards the creation of effective test inputs, but relatively little attention has been paid to the creation of oracles. The specification of test oracles remains challenging for many domains, such as real-time embedded systems, where small changes in timing or sensory input may cause large behavioral differences. Models of such systems, often built for analysis and simulation before the development of the final system, are appealing for reuse as oracles. These models, however, typically represent an idealized system, abstracting away certain considerations such as non-deterministic timing behavior and sensor noise. Thus, even with the same test data, the model’s behavior may fail to match an acceptable behavior of the SUT, leading to many false positives reported by the oracle. This talk will present an automated framework that can adjust, or steer, the behavior of the model to better match the behavior of the SUT in order to reduce the rate of false positives. This model steering is limited by a set of constraints (defining acceptable differences in behavior) and is based on a search process attempting to minimize a numeric dissimilarity metric. This framework allows non-deterministic, but bounded, behavior differences, while preventing future mismatches, by guiding the oracle—within limits—to match the execution of the SUT. Results show that steering significantly increases SUT-oracle conformance with minimal masking of real faults and, thus, has significant potential for reducing false positives and, consequently, development costs.
2:50 pm - 4:05 pm Friday, February 20 Swearingen 2A05	Dr. Jijun Tang Department of Computer Science & Engineering University of South Carolina	Title: Analyzing Genomes Beyond Sequences Abstract: Spatial conformation and interaction of chromatin play a fundamental role in important cellular functions.  With the influx of new data about the higher-level structure of genomes, new techniques are required to model, visualize and analyze the full extent of genomic information in three dimensions.  We recently developed the first model-view framework to integrate, model and visualize human genome in 3-dimension.  Here we report a new method on 3D genome model construction and a prototype of cloud based Genome3D viewer.  Our viewer uses high performance, cross-platform game engine that enables efficient rendering of human genome model in 3-dimension on Windows and Mac platform, which demands less computing power and makes 3D structural genome information available to a broad research community through web browser.  The integrated model-view framework could significantly advance genome research in integrating epigenomic data, studying long range inter- and intra-chromosome interaction, and analyzing structural features of genetic variations.
2:50 pm - 4:05 pm Friday, February 27 Swearingen 2A05	Dr. Srihari Nelakuditi Department of Computer Science & Engineering University of South Carolina	Title: Overview of Mobile Computing and Networking Research at USC
2:50 pm - 4:05 pm Friday, March 6 Swearingen 2A05	Dr. Jason Bakos Department of Computer Science & Engineering University of South Carolina	Title: Heterogeneous Computing at USC Abstract: The Heterogeneous and Reconfigurable Computing Lab develops tools and methodologies to make emerging processing technologies practical for high performance and embedded computing.  These technologies include massively data parallel processors, field programmable gate arrays, and automata processors.  We begin with an overview of these technologies, why they fill a crucial role in the future of computing, describe the efficiency metrics we use for their evaluation, and show examples of specific platforms we use in the lab.  Next we'll list and describe some of the problems we target and the results achieved.  Examples of these include the acceleration of applications in computational biology, data mining, linear algebra, computer vision, and graph algorithms.
2:50 pm - 4:05 pm Friday, March 20 Swearingen 2A05	Dr. Csilla Farkas Department of Computer Science & Engineering University of South Carolina	Title: Big Data Analytics: Privacy Protection using Semantic Web Technologies Abstract: In this talk I address the privacy issues in the context of big data analytics. The need to support data driven decision making led to the development of sophisticated technologies to collect and analyze large data sets. However, this useful means of analyzing data comes at a cost; sensitive data may be disclosed and individuals' privacy may be violated.
2:50 pm - 4:05 pm Friday, March 27 Swearingen 2A05	Dr. Gabriel Terejanu Department of Computer Science & Engineering University of South Carolina	Title: Towards Experimental Design Strategies for Inadequate Models Abstract. Obtaining informative measurements is a fundamental problem when inadequate models are used to guide the design of experiments. The focus of this study is to develop a basic understanding of the impact that modeling errors have on experimental design strategies. Through a rigorous modeling of structural errors, new adaptive experimental design strategies can be obtained by exploiting structural uncertainty. The feasibility of the proposed methodology is demonstrated in the context of contaminant dispersion models.
2:50 pm - 4:05 pm Friday, April 3 Swearingen 2A05	Dr. Xiaofeng Wang Depart of Electrical Engineering University of South Carolina	Title: Fault-Tolerant Control of Cyber-Physical Systems Abstract: As the complexity of Cyber-Physical Systems (CPS) increases, it becomes more and more challenging to ensure the reliability of CPS, especially in the presence of system failures. Simplex architecture is shown to be an efficient tool to address the software failure in such systems.  However, when physical failures also appear, Simplex does not work any more because the physical dynamics change due to physical failures.  The Simplex architecture designed for the original physical model may not be suitable for the new dynamics.  To address both software and physical failures, this talk presents the L1Simplex architecture, which contains the safety monitor, the high-performance controller (HPC), the L1-based high-assurance controller (HAC), and the decision logic for controller switching.  The safety monitor is used to monitor the system behavior.  It leads to another controller switching rule besides the stability-envelope-based rule in the decision logic.  The HAC is designed based on the L1 adaptive controller, with which the stability envelope is computed. We show that the L1Simplex architecture can efficiently handle a class of software and physical failures.
2:50 pm - 4:05 pm Friday, April 10 Swearingen 2A05	Dr. Alberto Quattrini Li Department of Computer Science, Electrical Engineering,  and Bioengineering Politecnico di Milano (Italy)	Title: On the Study, Design, and Evaluation of Exploration Strategies for Autonomous Mobile Robots Abstract: Exploration of initially unknown environments through the deployment of multirobot systems is an effective technique for many real-life applications, including map building and search and rescue. One of the most important and challenging aspects that could significantly impact on the system autonomy and performance is the decision about where to go next (exploration strategy) and about which robot goes where (coordination method) according to current knowledge of the environment. In this talk, I will present some results that contribute to the study, the design, and the evaluation of some aspects of such a decision-making process. In the first part, I will show a method for studying the optimal behavior obtainable by an exploring robot with limited and discrete visibility in a given environment represented as a grid. Further, I will present a worst- and average-case analysis of some exploration strategies used in practice on a graph-based environment. In the second part, I will present a multirobot exploration system based on semantic information, which contributes to improve the online exploration performance. In the third part, I will discuss about some aspects that could improve the experimental assessment of multirobot exploration systems, specifically by calculating the competitive ratio of some online exploration strategies and by systematically assessing some important factors affecting the exploration process, through repeatable experiments. The long-term endeavor is to contribute to make robots more efficient and autonomous, by shifting from ‘how to go there?’ to ‘where to go?’ paradigm.
3:30 pm – 4:30 pm Friday, April 17 Amoco Hall (Swearingen 1C01)	John Hodgson Blizzard Entertainment	Title: Games? Serious Games? Abstract: There are “games” (some of which are “computer games”), and there are “serious games.” What is meant by these terms? And are they the right terms? We will talk about why the term “serious game” might be a bad term. We will discuss how the values and purpose of video games informs their design and how the work of computer scientists and media artists can converge in the design and production of videogames.
2:50 pm - 4:05 pm Friday, April 24 Swearingen 2A05	Cancelled
Bonus Seminar 4 pm – 5 pm Friday, April 24 Swearingen 1A03 (Faculty Lounge)	Dr. Ioannis Rekleitis Department of Computer Science & Engineering University of South Carolina	Title: Autonomous Operations of Robots in the Field. Abstract: The last few years, robots have moved from the pages of science fiction books into our everyday reality. Currently, robots are used in scientific exploration, manufacturing, entertainment, and household maintenance. While the above advances were made possible by recent improvements in sensors, actuators, and computing elements, the research of today is focused on the computational aspects of robotics. This talk presents an overview of algorithmic problems related to robotics, with the particular focus on increasing the autonomy of robotic systems in challenging environments. In particular I would discuss the use of discrete structures as such as graphs to efficiently solve robotic problems. Cooperative Localization Mapping and Exploration employs teams of robots in order to construct accurate representations of the environment and of the robot's pose. The problem of coverage has found applications ranging from vacuum cleaning to humanitarian mine removal. A family of algorithms will be presented that solve the coverage problem efficiently in terms of distance travelled. I would present ongoing work on underwater robotics together with recent results from a multi-robot experiment employing a UAV, a USV, and an AUV operating in sync with a remote marine biologist located thousands of kilometers away. Finally, I will present some current work on the problem of searching under uncertainty. The work that I will present has a strong algorithmic flavour, while it is validated in real hardware.