Abstract: Traditional computer architecture primarily leverages abstracted interfaces such as instruction set architecture (ISA) and virtual memory mechanism to convey an application's information to the hardware. However, as pointed out in the community white paper “21st Century Computer Architecture”, such conventional interfaces are insufficient to convey more high-level requirement of applications to the hardware such as quality-of-service (QoS) and security, which are extremely important to data centers in the cloud era.
We propose a new computer architecture PARD (Programmable Architecture for Resourcing-on-Demand) that provides a new programming interface to enable more software-defined functionalities. PARD is inspired by the perspective that a computer is inherently a network in which hardware components communicate via packets (e.g., over the NoC or PCIe). Thus we can apply networking technologies, e.g. software-defined networking (SDN), to this intra-computer network.
In this talk, I will present an FPGA-based PARD prototype to show how to reconstruct a computer to be an SDN-like network, which enables new functionalities like fully hardware-supported virtualization and programmable application-specific QoS. Additionally, I will present ongoing work, i.e., QoS-aware data center software stack, including hypervisors, OS kernels and cluster management systems.
Bio: Yungang Bao is a Professor of Institute of Computing Technology (ICT), Chinese Academy of Sciences and serves as associate director of Center of Advanced Computer Systems of ICT. His research interests include computer architecture and operating systems and recently focuses on resource-efficient data center design. He co-led the Hybrid Memory Trace Tool (HMTT) project that has freely provided terabytes of off-chip memory traces of real systems to architecture community. During 2010-2012, he was a postdoc fellowship at Princeton University, working on the PARSEC 3.0 project. His work is published at many top venues such as ASPLOS, HPCA, ISCA and SIGMETRICS. He was the winner of CCF-Intel Young Faculty Researcher Program of the year for 2013 and was invited to participate in Dagstuhl Seminar on Rack-scale Computing in 2015. He received his Ph.D. degree from ICT (2008) and B.S. degree from Nanjing University (2003).
Abstract:Robots are increasingly leaving the well defined and controlled settings of factories and similar environments and they are entering all domains of our daily life. This dramatic change goes hand in hand with increasing demands for robotic systems that are capable of engaging in complex tasks in unstructured environments without a permanent supervision by human operators, i.e., these systems have to be able to perform feats that are attributed to cognition. These robots have to be able to perceive, i.e., to recognize objects or situations, to generate world models, i.e., to learn and maintain maps of their environment and of the items of interest therein, and to use these for executing their missions, often in close cooperation with humans or fellow robots. The talk gives on overview of according research activities of the Robotics Group at Jacobs University, especially in the context of logistics, search and rescue, and marine robotics. Jacobs University is a private research university founded in 1999 in Bremen, Germany. It has an international student body with ca. 1400 students from over 100 nations, admitted in a highly selective process.
Bio: Andreas Birk is a full professor (full) in Electrical Engineering and Computer Science at Jacobs University Bremen where he leads the robotics group. He started at Jacobs University in Fall 2001 while rejecting an offer for a professorship (C3) at the University of Rostock. Before he joined Jacobs University, he held a research-mandate of the Flemish Society for Applied Research, IWT. He was in addition from October 1997 on a visiting professor (docent) at the Vrije Universiteit Brussel (VUB).He also worked as a visiting professor (C3) at the Universit?t Koblenz-Landau in the winter-semester of 1999/2000. During the almost six years at the VUB, Andreas Birk was a member of the Artificial Intelligence Lab, which he joined as Postdoc in April 1996. In 1995 he received his doctorate from the Universit?t des Saarlandes, Saarbrücken, where he previously studied Computer Science from fall 1989 to spring 1993.
Head of the research lab for Autonomous Intelligent Systems
University of Freiburg
Abstract: Probabilistic approaches have been discovered as one of the most powerful approaches to highly relevant problems in mobile robotics including perception and robot state estimation. Major challenges in the context of probabilistic algorithms for mobile robot navigation lie in the questions of how to deal with highly complex state estimation problems and how to control the robot so that it efficiently carries out its task. In this talk, I will present recently developed techniques for efficiently learning a map of an unknown environment with a mobile robot. I will also describe how this state estimation problem can be solved more effectively by actively controlling the robot. For all algorithms I will present experimental results that have been obtained with mobile robots in real-world environments.
Bio: Prof. Dr. Wolfram Burgard is a professor for computer science at the University of Freiburg and head of the research lab for Autonomous Intelligent Systems. His areas of interest lie in artificial intelligence and mobile robots. His research mainly focuses on the development of robust and adaptive techniques for state estimation and control. Over the past years my group and I have developed a series of innovative probabilistic techniques for robot navigation and control. They cover different aspects such as localization, map-building, SLAM, path-planning, exploration, and several other aspects.
In his previous position from 1996 to 1999 at the University of Bonn he was head of the research lab for Autonomous Mobile Systems. In 1997 they deployed Rhino as the first interactive mobile tour-guide robot in the Deutsches Museum Bonn in Germany (see corresponding overview article). In 1998 his group and I went to Washington, DC, to install the mobile robot Minerva in the Smithsonian Museum of American History. Afterwards we produced several robots that autonomously operated in trade shows and Museums. In 2008, they developed an approach that allowed a car to autonomously navigate through a complex parking garage and park itself. In 2012, they developed the robot Obelix that autonomously navigated like a pedestrian from the campus of the Faculty of Engineering to the city center of Freiburg. He have published over 250 papers and articles in robotic and artificial intelligence conferences and journals. In 2005, I co-authored two books. Whereas the first one, entitled Principles of Robot Motion - Theory, Algorithms, and Implementations, is about sensor-based planning, stochastic planning, localization, mapping, and motion planning, the second one, entitled Probabilistic Robotics, covers robot perception and control in the face of uncertainty.
Dean of the School of Computer Science and Technology and School of Software Engineering, Shandong University.
Abstract:3D modeling of urban environments starts to play an increasingly important role in the emerging technologies from self-driving car to augmented reality. Beyond helping a human or a vehicle navigate, 3D urban models provide a base for spatially registering otherwise chaotic urban data, both sensor sensed and user generated, for better ‘mapping’ of urban big data. In this talk, I will introduce our decade long effort on acquiring and modeling large urban environments as well as analyzing and visualizing urban activities. I will also discuss future developments in this direction.
Bio: Baoquan Chen is endowed Changjiang Professor and Dean (CS & Software) of Shandong University. Prior to the current post, he was the founding director of the Visual Computer Research Center, Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences, and a faculty member at CS&E at the University of Minnesota at Twin Cities. His research interests generally lie in computer graphics, visualization, and human-computer interaction. He has published over 100 papers in top venues including ACM TOG/SIGGRAPH and IEEE TVCG.
Chen serves as associate editor of IEEE TVCG and on the steering committee of IEEE VIS. In the past, he served as conference chair of SIGGRAPH Asia 2014 and IEEE Visualization 2005, and program chair of SIGGRAPH Asia 2013 and IEEE Visualization 2004. Chen received PhD in CS from SUNY@Stony Brook, and MS in EE from Tsinghua. He received NSF CAREER award in 2003 and IEEE Visualization Best Paper Award in 2005.
Abstract:Realtime data processing powers many use cases at Facebook, including realtime reporting of the aggregated, anonymized voice of Facebook users, analytics for mobile applications, and insights for Facebook page administrators. We have a realtime data processing ecosystem at Facebook that handles hundreds of Gigabytes per second across hundreds of data pipelines. Many decisions must be made while designing a realtime stream processing system. In this talk, we identify five important design decisions that affect their ease of use, performance, fault tolerance, scalability, and correctness. We compare the alternative choices for each decision and contrast what we built at Facebook to other published systems. We will illustrate how our decisions and systems satisfy our requirements for multiple use cases at Facebook. Finally, we reflect on the lessons we learned as we built and operated these systems.
Bio: Jerry Chen is a software engineer at Facebook. He initiated and leads the effort to build Stylus, a high performance, scalable and fault tolerant stream processing system at Facebook. This system has been the enabler for many mission critical stream processing applications, such as Mobile Analytics, Instagram Trending, Page Insights, Chorus, etc. Before that, he managed the HBase and HDFS team at Facebook. Under his lead, HBase grew from an experimental project into a critical storage system powering Messages, as well as search index, operational datastore, etc. And HDFS team is powering one of the world's largest Hadoop clusters. Jerry Chen got his MSEE degree from University of Minnesota.
Abstract:Recently, human Robot Hybrid systems have been designed and developed to provide functional motion assistance to disabled and elderly people in daily activities. This talk will discuss key techniques of human-robot hybrid systems, which include ergonomics, physical human-robot interaction, wearable computing, human intention estimation, multimodal interaction and cooperation. The related advances in UESTC exoskeletons will also be introduced in this talk, which includes reinforcement learning in pHRI and our exoskeleton systems, AIDER system for walking assistance and HUALEX system for human augmentation.
Bio: Hong Cheng is a full professor of University of Electronic Science and Technology of China (UESTC), school of Automation and Engineering. He serves as an executive director of the Center for Robotics since 2014. He was a visiting scholar at School of Computer Science, Carnegie Mellon University, USA from 2006 to 2009. Before this, he received his Ph.D degree in Pattern Recognition and Intelligent Systems from Xi'an Jiaotong University in 2003 and became an associate Professor of Xi'an Jiaotong University since 2005. He joined UESTC since 2010.
His current research interests include machine learning in human robot hybrid systems. Prof. Cheng has over 80 academic publications including three books-“Digital Signal Processing (Tsinghua University Press, Sep. 2007)”, “Autonomous Intelligent Vehicles: Theory, Algorithms and Implementation (Springer, Dec. He served/is serving as a General Chair of VALSE 2015,Program Chair of CCPR 2016, and a General Chair for CCSR 2016 He is a senior member of IEEE.
Abstract: One of the first things a child learns is how to count. But counting in computer science is a very interesting problem with many applications and connections. One class of counting problems deals with Sum-of-Product computations. We will survey recent effort in classifying the computational difficulty of these counting problems. In line with the P vs. NP demarcation, we will describe dichotomy theorems that classify a broad class of counting problems, of which every problem is either computable in polynomial time (in the class P) or as difficult as it can be (#P-hard).
Bio: Jin-Yi Cai studied at Fudan University (class of 77). He continued his study at Temple University and at Cornell University, where he received his PhD in 1986. He held faculty positions at Yale University (1986-1989), Princeton University (1989-1993), and SUNY Buffalo (1993-2000), rising from Assistant Professor to Full Professor in 1996. He is currently a Professor of Computer Science and Steenbock Professor of Mathematical Sciences at the University of Wisconsin-Madison.
Dr.Cai received a Presidential Young Investigator Award in 1990, an Alfred P. Sloan Fellowship in Computer Science in 1994, a John Simon Guggenheim Fellowship in 1998, and a Morningside Silver Medal of Mathematics in 2004. He also received the Humboldt Research Award for Senior U.S. Scientists. He has been elected a Fellow of ACM and AAAS.
Abstract: Intelligent wheelchair is a typical mobility assistive robot designed to assist a user with a physical disability or cognitive impairments. Intelligent wheelchairs are interacting closely with humans and performing navigation tasks in human environments with unpredictable changes, the autonomy and safety issues are more essential for these complex and challenging situations. In this talk, I will introduce our recent work on intelligent wheelchair with emphasis on autonomy and human-safety in the aspects of mapping, localization, navigation and human-robot interaction. The prototype systems of intelligent wheelchair developed in our lab, as well as experimental studies in real and dynamic environments will be presented for illustrating our methodologies and applications."
Bio: Weidong Chen received his B.S. and M.S. degrees in Control Engineering in 1990 and 1993, and Ph.D. degree in Mechatronics in 1996, respectively, all from the Harbin Institute of Technology, Harbin, China. Since 1996, he has been at the Shanghai Jiao Tong University where he is currently Chair and Professor of the Department of Automation, and Director of the Institute of Robotics and Intelligent Processing. He is the founder of the Autonomous Robot Laboratory. From August 2003 to February 2004, he was a visiting associate professor in the Department of Electrical and Computer Engineering at The Ohio State University. In July and August 2012, he was a visiting professor in the Artificial Intelligence Laboratory at the University of Zurich in Switzerland.
Dr. Chen is a winner of the Shanghai Science and Technology Progress Award presented by the Shanghai Municipal Government in 2008, a co-winner of KUKA Service Robotics Best Paper Award of the IEEE International Conference on Robotics and Automation (ICRA) in 2011, a co-winner of the Award for Interactive Papers of the Combined 48th IEEE Conference on Decision and Control (CDC) and 28th Chinese Control Conference (CCC) in 2009, a winner of the Teaching Achievement Award presented by Shanghai Municipal Education Commission in 2004. He has also been recognized by the 2007 New Century Excellent Talents in University of Ministry of Education of China, and a 2004 Shanghai Shuguang Scholar. He is a co-founder of the JiaoLong RoboCup Team. JiaoLong RoboCup Team has won 10 national championships in Soccer Middle Size League and RoboCup@home League in the passed annual China Robot Competitions & China RoboCup Open since 2002.
Abstract: Remote Direct Memory Access (RDMA), a fast cross-machine memory access technique commonly seen in high-performance computing area, has recently gained increasing momentum in datacenter computing. In this talk, I will describe our recent efforts in leveraging RDMA to build efficient in-memory computing systems. Specifically, I will illustrate how to combine RDMA with hardware transactional.
Bio: Prof. Haibo Chen have been a Professor in School of Software, Shanghai Jao Tong University since December, 2001. He lead the Institue of Parallel and Distributed Systems and work memebers to improve the performance and dependablilty of computer systems. He received a Bachelor degree and a Ph.D degree (Advisor:Prof. Binyu Zang) in computer science, both from Fudan University.
Abstract: Developing efficient and scalable algorithms for Latent Dirichlet Allocation (LDA) is of wide interest for many applications. In this talk we use WarpLDA as an example to show how algorithm and system researchers can work together to build more efficient solutions. WarpLDA achieves O(1) time complexity per-token and fits the randomly accessed memory per- document in the L3 cache. Our empirical results in a wide range of testing conditions demonstrate that WarpLDA is consistently 5-15x faster than the state-of-the-art MH-based LightLDA, and is faster than the state-of-the-art sparsity aware F+LDA in most settings. WarpLDA manages to learn up to one million topics from billion-scale documents in 5 hours, at an unprecedentedly throughput of 11G tokens per second."
Bio: My research interest is parallel computing. In particular, I try to solve the problems of programming wall, memory wall, debugging wall and availability wall we face in parallel computing. In the recent two years, I have also been interested in mobile computing. My research in mobile computing aims at using cloud to manage operating systems, applications and data of mobile devices, and address the reliability, power and network bandwidth issues of current mobile computing platforms
Abstract: In this talk, we present a new particle-based approach to incompressible fluid simulation. We depart from previous Lagrangian methods by considering fluid particles no longer purely as material points, but also as volumetric parcels that partition the fluid domain. The fluid motion is described as a time series of well-shaped power diagrams (hence the name power particles), offering evenly spaced particles and accurate pressure computations. As a result, we circumvent the typical excess damping arising from kernel-based evaluations of internal forces or density without having recourse to auxiliary Eulerian grids. The versatility of our approach is demonstrated by the simulation of multiphase flows and free surfaces.
Bio: Mathieu Desbrun is the John W. and Herberta M. Miles Professor of Computing + Mathematical Sciences at the California Institute of Technology (Caltech). He regularly participates in a number of international program committees and editorial boards in computer graphics, including ACM SIGGRAPH and Eurographics. He now runs the Applied Geometry lab, focusing on discrete differential modeling, i.e., the development of differential, yet readily discretizable foundations for computational modeling. His research group has focused on a wide spectrum of applications, ranging from discrete geometry processing to solid and fluid mechanics and field theory, and discrete exterior calculus. Visit www.geometry.caltech.edu for more information.
Abstract: Currently unmanned aircraft are built from scratch, by adding custom autopilot and servomotors to the airframe from design. However, there are many cases that EXISTING aircraft are desired to fly unmanned to perform dull, dirty, and dangerous at moments’ notice. For example, during the Fukushima nuclear reactor accidents, helicopters are brought to pour water from above, but their safety was at great risk due to the unknown, far underestimated then, radiation coming from exposed core. If we can have a robot that pilot an aircraft, it will be able to make it fly from cold-start to parking Humanoid is by definition is a perfect candidate to do the job since it can nicely fit into existing cockpit, and can manipulate the various control knobs, dials, joysticks and more just like humans do. Further, it can even communicate with the aircraft onboard avionics using any available network so that it can process much more information at much faster way. Such robots can be also used not only in an airplane but in helicopters, automobile, or heavy equipment such as cranes or excavators. In this talk, a pioneering design of Pibot developed in KAIST, South Korea is presented. Its design concept and actual operation is showcased. We hope, someday, such robots will be used to help or even substitute human pilots if situation dictates.
Bio: Prof. David Hyunchul Shim is Associate Professor at KAIST, South Korea. Prof. Shim began his research on autonomous vehicles since 1991 and has been recognized as one of the pioneers in unmanned aerial vehicles for his work on flight control system design, collision avoidance, vision-based aerial mapping and navigation, and anti-drone technologies. He has developed a number of unmanned aerial vehicles including crop-dusting helicopters, tail-sitters, micro aerial vehicles and transforming aerial vehicles. He has coauthored more than 47 journal papers and more than 180 conference papers so far and cited more than 2,700 times by peers. Dr. Shim received the B.S. and M.S. degrees in mechanical design and production engineering from Seoul National University, Seoul, Korea, in 1991 and 1993, respectively, and the Ph.D. degree in mechanical engineering from the University of California Berkeley, Berkeley, USA in 2000. From 1993 to 1994, he was with Hyundai Motor Company, Korea, as Transmission Design Engineer. From 2001 to 2005, he was with Maxtor Corporation, Milpitas, CA, USA as Staff Engineer working on advanced control systems for hard disk drives. From 2005 to 2007, he was with the University of California Berkeley as Principal Engineer, in charge of Berkeley Aerobot Team. In 2007, he joined the Department of Aerospace Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejon, Korea, as an Assistant Professor. He is now a tenured Associate Professor and served as Director of Center of Field Robotics in KAIST Institute from ’12-’15. He is also serving as member and editor in several domestic and international journals and academic societies. From 2013, he has led the project planning team for nation-wide unmanned vehicles with Korean government. Now he is leading the entire project team for Korean civil UAV program. He is also serving as a consultant to a number of government committees for UAVs and autonomous cars research in Korea. He is also serving as a government-appointed advisor for Remotely Piloted Aircraft System Panel(RPAS/P) in International Civil Aviation Organization(ICAO) since 2015. He is also serving as an expert for AI & Robotics group in World Economic Forum. He has received a number of awards including the 2nd prize in Global Student Design Competition by National Instrument, USA(2014), Outstanding award by the Minister of Science, ICT and Future Planning of Korean Government, Best paper awards by Qualcomm Inc, Korean Aerospace Industry(’09,’14), and many more outstanding paper awards.
Abstract (Tutorial): Secure multi-party computation enables two (or more) mutually distributing parties to cooperatively compute a function on their private inputs, obtaining the output of the function but keeping both parties inputs and all intermediate values secret. Andrew Yao introduced the concept and a generic protocol for multi-party omputations in the 1980s, but it has only been in the past decade that it has become possible to build interesting applications using these techniques. In this tutorial, I will use secure stable matching as a motivating problem to introduce multi-party computation (without assuming prior background or extensive knowledge of cryptography) and discuss some of the techniques that have been develop to enable over five orders of magnitude performance improvements in MPC executions over the past decade.
Abstract (Symposium): Hiding memory access patterns is required for secure computation, but remains prohibitively expensive for many interesting applications. Prior work has either developed custom algorithms that minimize the need for data-dependant memory access, or proposed the use of Oblivious RAM (ORAM) to provide a general-purpose solution. However, most ORAMs are designed for client-server scenarios, and although they provide asymptotic performance benefits, because of their high initialization and concrete costs they remain worse than linear scan for most actual uses.
In this talk, I will present work on a new ORAM approach for secure computation that revisits the classical square-root ORAM design of Goldreich and Ostrovsky. Our design replaces the expensive pseudo-random function needed in traditional hierarchical ORAM designs with inexpensive permutations and a recursive position map, and achieves orders of magnitude performance improvements over the best previous ORAM designs for use in secure computation applications. By combining the ORAM design with custom data structures, we are able to implement previously infeasible algorithms, including secure stable matching at the scale needed for national medical residency matching, as secure computations.
This talk covers joint work with Samee Zahur, Jack Doerner, Xiao Wang, Mariana Raykova, Adrià Gascón, Jonathan Katz, and abhi shelat.
Bio: David Evans is a Professor of Computer Science at the University of Virginia and leader of the Security Research Group. His research focuses on privacy and security for computing systems, and empowering individuals and organizations to control how their data is used and shared. He is the author of an open computer science textbook, a children's book on combinatorics and computability, and teacher of one of the world's most popular MOOCs. He won the Outstanding Faculty Award from the State Council of Higher Education for Virginia and was Program Co-Chair for the 31st and 32nd IEEE Symposia on Security and Privacy. He has SB, SM and PhD degrees in Computer Science from MIT and has been a faculty member at the University of Virginia since 1999.
Abstract: Multi-view image-based 3D reconstruction techniques have changed the world in the last ten years. Researchers competed in quantitative evaluation studies (e.g., Middlebury benchmark) around 2006. The technology had made dramatic progress in a few years and became mature enough to be deployed in various real products in industry. With state-of-the-art techniques, one can reconstruct a dense mesh model of an object by just waving a cell phone, or city-scale building models from hundreds of thousands of aerial photographs fully automatically. Computer Vision "solved" image-based dense geometry reconstruction in a sense. I will give a talk on what are the next challenges and a series of research from my group towards these goals.
Bio: I am an assistant professor of Computer Science and Engineeering at Washington University in St. Louis. Prior to WUSTL, I was a software engineer at Google. Before Google, I was a post-doctoral research associate at University of Washington. I worked with Prof. Seitz and Prof. Curless at University of Washington, and Rick Szeliski at Facebook (was at Microsoft Research). I completed my Ph.D. under the supervision of Prof. Ponce at Computer Science Department of Univeristy of Illinois at Urbana-Champaign in May 2008.
Abstract: Immersive visual and experiential computing systems, i.e. virtual and augmented reality (VR/AR), are entering the consumer market and have the potential to profoundly impact our society. Applications of these systems range from communication, entertainment, education, collaborative work, simulation and training to telesurgery, phobia treatment, and basic vision research. In every immersive experience, the primary interface between the user and the digital world is the near-eye display. Thus, developing near-eye display systems that provide a high-quality user experience is of the utmost importance. Many characteristics of near-eye displays that define the quality of an experience, such as resolution, refresh rate, contrast, and field of view, have been significantly improved over the last years. However, a significant source of visual discomfort prevails: the vergence-accommodation conflict (VAC). Further, natural focus cues are not supported by any existing near-eye display. In this talk, we discuss frontiers of engineering next-generation opto-computational near-eye displays systems to increase visual comfort and provide realistic and effective visual experiences.
Bio: Prior to joining Stanford University's Electrical Engineering Department as an Assistant Professor in 2014, Gordon Wetzstein was a Research Scientist in the Camera Culture Group at the MIT Media Lab. His research focuses on computational imaging and display systems as well as computational light transport. At the intersection of computer graphics, machine vision, optics, scientific computing, and perception, this research has a wide range of applications in next-generation consumer electronics, scientific imaging, human-computer interaction, remote sensing, and many other areas. Gordon's cross-disciplinary approach to research has been funded by DARPA, NSF, Samsung, Intel, and other grants from industry sponsors and research councils. In 2006, Gordon graduated with Honors from the Bauhaus in Weimar, Germany, and he received a Ph.D. in Computer Science from the University of British Columbia in 2011. His doctoral dissertation focuses on computational light modulation for image acquisition and display and won the Alain Fournier Ph.D. Dissertation Annual Award. He organized the IEEE 2012 and 2013 International Workshops on Computational Cameras and Displays, founded displayblocks.org as a forum for sharing computational display design instructions with the DIY community, and presented a number of courses on Computational Displays and Computational Photography at ACM SIGGRAPH. Gordon won best paper awards at the International Conference on Computational Photography (ICCP) in 2011 and 2014 as well as a Laval Virtual Award in 2005.
Abstract: With the advancement of cloud computing, big data and artificial intelligence, internet video becomes a very fast growing service. This talk will introduce the fast development of internet videos and the big challenge of object detection and virtual reality for internet video services. We'll present the research in CNN based object detection, which is used for content based ads recommendation. Then we will talk about new advancement and viewpoints in recent virtual reality practice.
Bio: Dr. Tao Wang, Distinct member of China Computer Federation (CCF), Vice director of Technical Committee of Computer Vision, CCF. He is chief scientist of iQIYI ltd. Corp., the biggest video sharing platform in China, where he works on computer vision and multimedia software applications. He received his Ph.D. in computer science from Tsinghua University in 2003. Tao then worked as a senior researcher in Intel Labs China during 2003~2014. He has published more than 60 papers in IJCV, CVPR, CIVR, ICME, and ACM multimedia etc.
Abstract: This presentation is designed to address the third wave of computing with a focus on immersive computing and pervasive machine intelligence. The discussion will showcase our interaction with computers and how that leads to new ways of communication and creativity. In turn, we are presented with new ways to visualize that creativity – specifically virtual reality. We must ask ourselves: what do we want to achieve with VR and where are going with VR. The ultimate goal is to reach full immersion and goes beyond just visuals alone -- all the senses need to be considered. To achieve this realization, we must understand and properly utilize computing power to accomplish these goals.
Bio: Raja Koduri is senior vice president and chief architect of the AMD Radeon Technologies Group, responsible for overseeing all aspects of graphics technologies used in AMD’s APU, discrete GPU, semi-custom, and GPU compute products.
Koduri brings more than 20 years of hands-on experience advancing the visual computing experience on a broad range of computing platforms, including PCs, game consoles, professional workstations and consumer devices, and across the content ecosystems. He most recently drove AMD’s innovation in visual and accelerated computing as corporate vice president of visual and perceptual computing, including the development of the industry’s first graphics chip with integrated High-Bandwidth Memory (HBM) and AMD’s LiquidVR? initiative, a series of technologies and toolkits to deliver more compelling and realistic VR experiences.
Abstract: Flying robots can operate in three-dimensional, indoor and outdoor environments. However, many challenges arise as we scale down the size of the robot, which is necessary for operating in cluttered environments. I will describe recent work in developing small, autonomous robots, and the design and algorithmic challenges in the areas of (a) control and planning, (b) state estimation and mapping, and (c) coordinating large teams of robots. I will also discuss applications to search and rescue, first response and precision farming. Publications and videos are available at kumarrobotics.org.
Bio: VIJAY KUMARis the Nemirovsky Family Dean of Penn Engineering with appointments in the Departments of Mechanical Engineering and Applied Mechanics, Computer and Information Science, and Electrical and Systems Engineeringat the University of Pennsylvania. Dr. Kumar received his Bachelor of Technology degree from the Indian Institute of Technology, Kanpur and his Ph.D. from The Ohio State University in 1987. He has been on the Faculty in the Department of Mechanical Engineering and Applied Mechanics with a secondary appointment in the Department of Computer and Information Science at the University of Pennsylvania since 1987.
Dr. Kumar served as the Deputy Dean for Research in the School of Engineering and Applied Science from 2000-2004. He directed the GRASP Laboratory, a multidisciplinary robotics and perception laboratory, from 1998-2004. He was the Chairman of the Department of Mechanical Engineering and Applied Mechanics from 2005-2008. He served as the Deputy Dean for Education in the School of Engineering and Applied Science from 2008-2012. He then served as the assistant director of robotics and cyber physical systems at the White House Office of Science and Technology Policy (2012 – 2013).
Dr. Kumar is a Fellow of the American Society of Mechanical Engineers (2003), a Fellow of the Institution of Electrical and Electronic Engineers (2005) and a member of the National Academy of Engineering (2013). Dr. Kumar’s research interests are in robotics, specifically multi-robot systems, and micro aerial vehicles. He has served on the editorial boards of the IEEE Transactions on Robotics and Automation, IEEE Transactions on Automation Science and Engineering, ASME Journal of Mechanical Design, the ASME Journal of Mechanisms and Robotics and the Springer Tract in Advanced Robotics (STAR).
Professor of Dept. of Electronic and Computer Engineering,UST,HK.
Professor of Automation Technology Cooperative Research Center,UST,HK.
Abstract: In this talk, I will describe our effort spanned over the last twenty some years to develop the so-called HKUST robotic ecosystem. This includes setting up a basic research program to study fundamental problems of robotics using differential geometric tools (geometric intuitions gained here are subsequently applied to cutting-edge technology and product developments), introducing project-based learning into our curriculum to empower students with critical hands-on and teamwork skills needed for design and rapid realization of complex robotic systems, and establishing mentoring and incubation programs to facilitate the often challenging and difficult process of students-based startups. A number of companies including Googol Tech, a leading motion control company, DJI, a world renowned drone manufacturer, QKM, a rising star in providing robotic solutions to the 3C industry, and ePropulsion, an innovative electric and environmentally friendly outboard products company, have spun off from HKUST robotics lab. Feedbacks from these spinoffs and close interactions with our industry partners provide critically needed data for continuousimprovement of our teaching and research programs.
Rapid prototyping and fast scaling ups are essential for any robotic startups to succeed. We benefited greatly from the amazingmanufacturing eco-system of the Pearl River Delta (PRD) region (also known as the Hollywood of Makers) and believe that this resource is also sought-after by robotic startups elsewhere. We established the SongshanLake Robotic Startup Facility (S^2L XbotPark) at the heart of the PRD region to assist entrepreneurs from the global robotics community. I will highlight some key features of the S^2L XbotPark.
Bio: Zexiang Li attended the South-Central University in 1978, received his BS (with honor) degrees in Electrical Engineering and Economics from Carnegie-Mellon University in 1983, his MS degree in EECS in 1985, MA in mathematics and PhD in EECS in 1989, all from the University of California at Berkeley. He worked at ALCOA, the Robotics Institute of CMU and the AI Lab of MIT (89-90). He was an assistant professor at the Courant Institute of New York University (90-92). In 1992, he joined the Department of Electronic and Computer Engineering of the Hong Kong University of Science and Technology and is currently a professor of the department. He founded the Automation Technology Center (ATC) and the Robotics Institute (RI) of HKUST.
Zexiang Li received the ALCOA Foundation Fellowship in 1979, and the E. Anthony Fellowship in 1983. He was a recipient of the University Scholar award from CMU in 1983, the E.I. Jury award from UC Berkeley in 1989, the Research Initiation award from NSF (US) in 1990, the Outstanding Young Researcher award (Class B) from NSF China in 2000, the LEAD award from AMI, USA in 2001, and the Natural Science award (3rd class) from China in 1997. He became an IEEE Fellow in 2008.
Zexiang Li served as a panel member of the Hong Kong Research Grants Council (RGC), an overseas member of the Natural Science Foundation of China (NSFC), and an associate editor for the IEEE Trans. on Robotics and Automation. He was the general Chair for the 2011 IEEE International Conference on Robotics and Automation (ICRA).
ZexiangLi's research areas of interests include multifingeredrobotic hand, parallel manipulators, workpiecelocalization and inspection, motion control, precision assembly, and unmanned aerial vehicles (UAVs). He is the author of more than 100 journal and conference papers, and four books, including A Mathematical Introduction to Robotic Manipulation (CRC Press 1993), and Nonholonomic Motion Planning (Kluwer 1994).
Zexiang Li co-founded several companies with his colleagues and students from the Automation Technology Center, including Googol Technology, a leading motion control company in China, DJI, a global leader in drones products, Lie Group Automation (or QKM), ePropulsion, the SongshanLake Robotic Startup Facility (S^2L XbotPark) and the Clearwater Bay Venture Capital.
Graduate School of Information Science and Technology, Osaka University.
Abstract: 3D modeling of real-world objects plays a central role in virtual reality systems. This talk introduces a photometric approach to high-quality 3D reconstruction from images, with which the shape of a target surface is recovered from the shading information under some illumination conditions. This talk mainly covers a photometric stereo method which uses observations under varying illumination conditions, starting from a basic least-squares solution method to more advanced robust methods for dealing with diverse reflectance properties.
Bio: Yasuyuki Matsushita received his B.S., M.S. and Ph.D. degrees in EECS from the University of Tokyo in 1998, 2000, and 2003, respectively. From April 2003 to March 2015, he was with Visual Computing group at Microsoft Research Asia. In April 2015, he joined Osaka University as a professor. His research area includes computer vision, machine learning and optimization. He is on the editorial board of IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), International Journal of Computer Vision (IJCV), IPSJ Journal of Computer Vision and Applications (CVA), The Visual Computer Journal, and Encyclopedia of Computer Vision. He served/is serving as a Program Co-Chair of PSIVT 2010, 3DIMPVT 2011, ACCV 2012, ICCV 2017, and a General Co-Chair for ACCV 2014. He is a senior member of IEEE.
Abstract: coming soon
Bio: Mr Raymond is responsible for VR technology, strategic partnership, VR business in Asia Pacific, as well as creates the exceptional product and brand value on HTC Vive. He has abundant experience and knowledge in consumer electronics and technology.
From 2006 to 2008, he worked with Google Android team and developed the World 1st Android Smartphone and was in charge of software development in Android platform.Graduated and received 2 Master Degrees in Mechanical Engineering from National Taiwan University and Electrical Engineering from University of Southern California.
Abstract: We believe that the scope of the research agenda of Cyber Physical Systems (CPS) has now expanded to include the Internet of Things (IoT), Mechanism and Incentive Design, Resilience and Cyber Security, and indeed data analytics for Big Data arising from CPS Systems. I would like to expand on this theme in this talk: There has been a great deal of work in recent years on the evolution of “Action Webs”. By this we mean closing the loop around IoT networked sensors. These networked control systems are fast becoming the next generation societal-scale. Cyber Physical Systems infrastructure for systems such as power, water, building systems, etc. Societal Scale CPS systems need to offer new data oriented service models, be robust, fault tolerant, and able to operate through cyber-attacks. Thus, the agenda of Societal Scale CPS Systems very much expands on IoT. Additionally, the advent of Societal Scale CPS Systems is causing the emergence of new models for monetization for the offering of new data oriented services. In this talk, we will provide the rudiments of a theory of resilient societal CPS systems, including the modeling of utility based privacy and security. Further, we will discuss how we can provide economic incentives to the private entities which own individual action webs to address the issues of “under investment in the common good”. More technically, this is a mechanism design procedure for helping bridge the gap between the non-cooperative Nash equilibrium of multiple players and the societal optimum strategy. Since we need these Societal Scale CPS systems to be “always on”, we need to build in the attributes of resilience and cyber security into these systems. Taxonomies of attack of Societal CPS systems are richer than those of traditional cybersecurity. Thus, we model the resilience of societal scale cyber physical systems to information attack, and how to operate them through successful attacks.
The talk is based on joint work with Lillian Ratliff, Roy Dong, Eric Mazumdar, Henrik Ohlsson (C3 Energy, LLC), Galina Schwartz, Saurabh Amin (MIT), and Alvaro Cardenas (UT Dallas).
Bio: Dr.Shankar Sastry is currently the Dean of Engineering at University of California, Berkeley and the faculty director of the Blum Center for Developing Economies. From 2004 to 2007 he was the Director of CITRIS (Center for Information Technology in the Interests of Society) an interdisciplinary center spanning UC Berkeley, Davis, Merced and Santa Cruz.
Dr.Shankar Sastry has served as Chairman, Department of Electrical Engineering and Computer Sciences, University of California, Berkeley from January, 2001 through June 2004. From 1999-early 2001, he was on leave from Berkeley as Director of the Information Technology Office at the Defense Advanced Research Projects Agency (DARPA). From 1996-1999, he was the Director of the Electronics Research Laboratory at Berkeley.
Abstract: Mapping is an important part of many robotic applications. Topological maps are abstractions that represent certain places as vertices and connections between those places as edges. Often Voronoi Diagrams are used to extract Topological maps from 2D grid maps. Since this does not work well in open areas, a room detection algorithm based on alpha shapes is presented, that extends the Topological map generation to represent for big open areas as a single vertex in the graph. Having two overlapping topological maps we can attempt to match those maps together. We show several similarity measures for those matches, including a new path similarity. The matched maps can then be used for different applications such as map merging, localization, planning or map evaluation.
At the end of the talk the newly formed ShanghaiTech Automation and Robotics Center (STAR-Center) of the School of Information Science and Technology (SIST) and some of its planned research activities will shortly be presented.
Bio: S?ren Schwertfeger received his Diploma (German equivalent of the Master) in Computer Science in 2005 from the University of Bremen in Germany. In 2012 he received his Ph.D. in Computer Science from the Jacobs University Bremen, Germany. Between 2012 and 2014 he was a postdoctoral researcher at the Robotics Group of Prof. Andreas Birk at the Jacobs University Bremen. In 2010 Dr. Schwertfeger was a guest researcher at the National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland, USA, where he did research on robot performance evaluation and map quality assessment.
His research interest is in robotics, especially in intelligent functions for mobile robots. Besides his work on map evaluation, Dr. Schwertfeger also worked and published on mapping, object detection and robot autonomy. He successfully participated in many robot competitions, both as team member and as judge and organizer. Dr. Schwertfeger was the chair of the 2014 German Open RoboCup Rescue competition. In 2012 he won the best paper award at the International Workshop on Safety, Security, and Rescue Robotics (SSRR). He is the reviewer for many top journals and conferences in the area of robotics.
Abstract: We give near optimal algorithms for regression, low rank approximation, and robust variants of these problems. Our results are based on the sketch and solve paradigm, which is a tool for quickly compressing a problem to a smaller version of itself, for which one can then run a slow algorithm on the smaller problem. These lead to the fastest known algorithms for fundamental machine learning and numerical linear algebra problems, which run in time proportional to the number of non-zero entries of the input. We first give algorithms for least squares regression, and robust variants such as l_p regression and M-Estimator loss functions. Then we give algorithms for approximate singular value decomposition, and robust variants such as minimizing sum of distances to a subspace, rather than sum of squared distances. Time-permitting, we also discuss communication-efficient solutions to these problems in distributed environments.
Bio: David Woodruff joined IBM Almaden Research Center in 2007 right after completing his Ph.D. at MIT in theoretical computer science. He has been at IBM Almaden ever since. His research interests include communication complexity, data stream algorithms, machine learning, numerical linear algebra, sketching, and sparse recovery. He is the recipient of the 2014 Presburger Award and Best Paper Awards at STOC 2013 and PODS 2010. At IBM he is a member of the Academy of Technology and a Master Inventor.
Abstract: In this talk, I will discuss about our recent research on smart data center infrastructures, especially data center power and networking. For networking, we will present our work on a 12-rack, 180-server DCN using multiwavelength switching and interconnection testbed. We implement real-time network traffic and per-link utilization monitoring, and full-stack optimization by jointly optimizing optical switching and network flow routing. We show that the data driven approach significantly improves both data center network and wide-area network performance. For data center power, we show that using probabilistic control based on workload data, we can allocate more machines into a data center given a fixed power capacity, without any performance disturbances to existing applications. The presentation is based on our recent publications on EuroSys, Sigcomm, OFC and other conferences and workshops.
Bio: Prof. Wei Xu is an assistant professor at the Institute for Interdisciplinary Information Sciences of Tsinghua University in Beijing. He is a recipient of the National Youth 1000 Program (青年千人计划) in 2013. He have a broad research interest in distributed system design and big data. My current projects include data center networking, system management and debugging, large scale system for machine learning and data mining, as well as various big data applications.He received my Ph.D from UC Berkeley in 2010. He was in the RAD Lab in EECS Department. My advisors are Prof. David Patterson and Prof. Armando Fox. His dissertation is on analyzing free text console logs for problem detection. He worked for Google for 2.5 years as a software engineer before joining Tsinghua University. He is the director of Open Compute Project (OCP) Certification Lab in China. I am also the director of international partnership for the MOE Research Center for Online Education.
Dean of Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University.
Distinguished Professor-at-Large, Chinese University of Hong Kong.
Abstract: Increasingly, the concepts and methods of computer science are being recognized as a source of great intellectual interest, injecting fresh ideas into other scientific disciplines. Through discourses and collaborations, exciting multidisciplinary areas are blossoming. We illustrate this phenomenon from the viewpoint of Theory of Computation.
Bio: Prof. Andrew Chi-Chih Yao, world-leading computer scientist, winner of the A.M. Turing Award in 2000. He is member of the US National Academy of Sciences and the American Academy of Arts and Sciences, foreign member of the Chinese Academy of Sciences, and fellow of the International Association for Cryptologic Research (IACR). He is also Dean of Institute for Interdisciplinary Information Sciences, Tsinghua University, Chair Professor of “Tsinghua Xuetang Special Pilot CS Class”, 973 Program Chief Scientist and Distinguished Professor-at-Large at the Chinese University of Hong Kong. Born in Shanghai in 1946, he received a BS in Physics from National Taiwan University in 1967, a PhD in Physics from Harvard University in 1972, and a PhD in Computer Science from the University of Illinois in 1975. After serving on the faculty at the Massachusetts Institute of Technology (1975-1976), Stanford University (1976-1981, 1982-1986) and the University of California at Berkeley (1981-1982), he joined Princeton University in 1986 as the William and Edna Macaleer Professor of Engineering and Applied Science. In 2004, he left Princeton to become a Professor at Tsinghua University and founded "Tsinghua University Special Pilot CS Class", Institute for Theoretical Computer Science, Institute for Interdisciplinary Information Sciences and Center for Quantum Information at Tsinghua University.
His research interests include theory of computation and its application in cryptography and quantum computing. Prof. Yao has made research contributions in three ways: (1) creating important subfields for theoretical computer science, (2) helping lay the foundations of modern cryptography, and (3) resolving open problems and establishing new paradigms in circuit complexity, computational geometry, data structures, and quantum computing. As a leading scientist in network communication complexity theory, Prof. Yao first developed the quantum communication complexity in 1993, which has laid the theoretical foundations for quantum computer. He developed the distributed quantum computation model in 1995, which has evolved into the basics of distributed quantum algorithms and quantum communication protocol security. Prof. Yao was awarded the A.M. Turing Award in 2000 for his contributions to the theory of computation, including the complexity-based theory of pseudorandom number generation, cryptography, and communication complexity. He is the first Asian laureate of the Turing Award since its establishment and the only Chinese laureate so far. The Turing Award is recognized as the "highest distinction in Computer Science" and the "Nobel Prize of computing". He has also received numerous other honors and awards including the George Polya Prize and the first Donald E. Knuth Prize, and several honorary degrees from University of Waterloo, the Chinese University of Hong Kong, the Hong Kong University of Science and Technology, and the City University of Hong Kong. In February 2009, he was elected as one of the “Capital Top Ten Education Newsmakers of 2009”.
Professor, School of Information Sciences and Technology, ShanghaiTech.
Professor, College of Engineering, University of Delaware.
Abstract: There is a huge gap between current VR content perceived through VR displays and real world content perceived naturally by human eyes. Human eyes perceive the 3D world through three cues: binocular stereo, monocular refocusing cues, and motion parallax. However, no VR content generated today can simultaneously reproduce all three effects. In this talk, Prof. Jingyi Yu will present a light field approach, developed by ShanghaiTech University and PlexVR, that enables the eyes to see VR content with all three effects. Our light field technologies cover light field acquisition, light field rendering, and light field display to provide unprecedented visual realism in VR and AR.
Bio: Jingyi Yu is a Full Professor in the School of Information Science and Technology at ShanghaiTech University. He received B.S. from Caltech in 2000 and Ph.D. from MIT in 2005. Before joining ShanghaiTech, he was a full professor at the University of Delaware. He has published over 100 papers at highly refereed conferences and journals including over 50 papers at the premiere conferences and journals CVPR/ICCV/ECCV/TPAMI. He has been granted 10 US patents on computational imaging. He is a recipient of the NSF CAREER Award and the US Air Force YIP Award. He is currently an Associate Editor of IEEE TPAMI, Elsevier CVIU, Springer TVCJ and Springer MVA. In 2016, he founded PlexVR, a startup that focuses on light field technologies for VR contents generation.
Bio: Zheng Zhang is professor of computer science at NYU Shanghai. He also holds an affiliated appointment with the Department of Computer Science at the Courant Institute of Mathematical Sciences and with the Center for Data Science at NYU's campus in New York City. Prior to joining NYU Shanghai, he was founder of the System Research Group in Microsoft Research Asia, where he served as principle researcher and research area manager. Before he moved to Beijing, he was project lead and member of technical staff in HP-Labs. He holds a PhD from University of Illinois, Urbana-Champaign, an MS from University of Texas, Dallas, and a BS Fudan University.
Zhang is a member of the Association for Computing Machinery and founder of the SIGOPS APSYS workshop and the CHINASYS research community. He served regularly as PC members of leading system conferences. During his tenures in industrial labs, he was awarded 40 patents and made numerous contributions to product lines. He has several Best Paper awards as well as awards for excellence from Microsoft and HP-Labs.