Special Session

Special Session
Special Session I: Systems & Applications

08:20-10:00, October 18 (Wednesday), 2017

Time Title Invited Speakers
08:20-10:00 Providing Usability Support for Mobile Applications in a Vehicular Environment Dr. Kyungmin Lee,
Facebook, USA
Mobile Deep Learning Framework for Continuous Vision Sensing Prof. Youngki Lee,
Singapore Management University, Singapore
Enabling Customized Data Management on Mobile Systems Prof. Steve Ko,
University at Buffalo, USA

Invited Talk 1: ”Providing Usability Support for Mobile Applications in a Vehicular Environment”
Dr. Kyungmin Lee, Facebook, USA

We live in a world where mobile computing systems are increasingly integrated with our day-to-day activities. Unlike the traditional desktop environment, people interact with mobile applications while performing other primary tasks such as walking and driving. As a result, developers can no longer assume mobile applications will always receive the user’s full attention. Furthermore, with the rise of wearable computing platforms such as Google Glass and smart watches, and with the deployment of pervasive platforms such as in-vehicle infotainment systems, the contention for available user attention will grow even more between mobile applications and the user’s primary activity.
One of the environments where the user has a limited amount of attention to interact with mobile applications is the vehicular environment. In this talk, I will present how I can provide system support for providing better usability in a vehicular environment. I will first discuss AMC (Android Model Checker), a toolkit that I have developed to assist application developers to automatically check their applications for vehicular UI guideline violations. Then, I will discuss my latest work, Gremlin, which requires the mobile OS to manage user attention in a vehicular setting. Gremlin ensures high usability of mobile applications since it only allows mobile applications to initiate an interaction when the user has sufficient amount of available attention to handle the interaction.

Kyungmin Lee has recently joined Facebook. Prior to joining Facebook, he was a researcher at IBM T.J. Watson Research Center. His research interest lies in the intersection of mobile systems and usability. More specifically, he has solved usability problems that arise from using mobile systems in a vehicular environment and when the user’s available attention is limited. He has also created mechanisms for providing high usability of mobile cloud-based gaming applications in spite of a high network latency. Moving forward, he is interested in providing system support to enable pervasive augmented reality (AR) & virtual reality (VR) experience in mobile devices. He received the Best Paper Award from the 13th International Conference on Mobile Systems, Applications, and Services (MobiSys ’15), Best Demo Award from MobiSys ’14, and runner-up for the Best Paper Award from MobiSys ’13. He received his Ph.D. in Computer Science and Engineering at the University of Michigan under the supervision of Professor Jason Flinn and Professor Brian Noble


Invited Talk 2: ”Mobile Deep Learning Framework for Continuous Vision Sensing”
Prof. Youngki Lee, Singapore Management University, Singapore

The rapid emergence of head-mounted devices such as the Microsoft Hololens enables a wide variety of continuous vision applications. Such applications often adopt deep-learning algorithms such as CNN and RNN to extract rich contextual information from the first-person-view video streams. Despite the high accuracy, use of deep learning algorithms in mobile devices raises critical challenges, i.e., high processing latency and power consumption. In this talk, I am going to introduce our early efforts to build a mobile deep learning inference system named DeepMon. DeepMon runs a variety of deep learning inferences purely on a mobile device in a fast and energy-efficient manner. For this, we designed a suite of optimization techniques to efficiently offload convolutional layers to mobile GPUs and accelerate the processing; note that the convolutional layers are the common performance bottleneck of many deep learning models. Our experimental results show that DeepMon can classify an image over the VGG-VeryDeep-16 deep learning model in 644ms on Samsung Galaxy S7, taking an important step towards continuous vision without imposing any privacy concerns nor networking cost.

Youngki Lee is an assistant professor at Singapore Management University since March 2013. He received the Ph.D. degree in Computer Science from KAIST. He has broad research interests in building experimental and creative software systems, which covers a wide design spectrum across operating systems, applications, and users. More specifically, his research interest lies in building underlying mobile and sensor platforms to enable always-available and highly-enriched awareness on human behavior and surrounding contexts. He is also interested in building innovative life-immersive mobile sensing applications in various domains such as daily well-being, childcare, and advertisement in collaboration with domain experts. More details about him can be found at http://youngkilee.blogspot.com.


Invited Talk 3: ”Enabling Customized Data Management on Mobile Systems”
Prof. Steve Ko, University at Buffalo, USA

In this talk, I will discuss a pluggable data management solution for modern mobile platforms (e.g., Android) that my research group is working on. Our goal is to allow data management mechanisms and policies to be implemented independently of core app logic. Our design allows a user to install data management solutions as apps, install multiple such solutions on a single device, and choose a suitable solution each for one or more apps. It allows app developers to focus their effort on app logic and helps the developers of data management solutions to achieve wider deployability. Finally, it gives increased control of data management to the end users and allows them to use different solutions for different apps. I will discuss our prototype implementation called BlueMountain, as well as several data management solutions we have implemented for file and database management to demonstrate the utility and ease of using our design.

Steve Ko is an associate professor in the Department of Computer Science and Engineering at the University at Buffalo, The State University of New York. He is generally interested in systems, and his current research focus is on mobile systems. Before joining UB, he graduated with PhD from UIUC and was a postdoc at Princeton. He is a recipient of the CAREER Award from NSF in 2014, the Young Investigator Award from UB in 2014, the Engineering Teacher of the Year Award from UB Engineering in 2015, the Teaching Innovation Award from UB in 2016, and the Distinguished Alumni Educator Award from UIUC in 2017.

Special Session II: ITS & V2X

16:00-17:40, October 18 (Wednesday), 2017

Time Title Invited Speakers
16:00-17:40 Turbo Controller Area Network (Turbo CAN): Next Generation In-Vehicle Networking Supporting +100Mbps and Being Compatible with CAN Prof. Ji-Woong Choi,
DGIST, Korea
NTU-NXP Smart Mobility Test-Bed: A Campus-Wide Infrastructure for Connected Cars Dr. Yong Liang Guan,
NTU Singapore, Singapore
V2X Channel Characteristics: Guidelines for Designing V2X Service Applications Prof. Dong Seog Han,
KNU, Korea

Invited Talk 1: ”Turbo Controller Area Network (Turbo CAN): Next Generation In-Vehicle Networking Supporting +100Mbps and Being Compatible with CAN”
Prof. Ji-Woong Choi, DGIST, Korea

As the number of electronic components in the car increases, the requirement for the higher data transmission scheme among them is on the sharp rise. The controller area network (CAN) has been widely adopted to support the in-vehicle communications needs but the data rate is far below what other schemes such as Ethernet and optical fibers can offer. A new scheme, named Turbo CAN, for enhancing the speed of CANs has been proposed recently, where a carrier modulated signal is introduced on top of the existing CAN signal, whereby the data rate can be enhanced over 100 Mbps. Turbo CAN is compatible with the existing CAN network and accordingly enables seamless upgrade of the existing network to support high-speed demand using CAN protocol and easier interfacing with vehicle to something (V2X) data. In this talk, I will briefly explain the trends of in-vehicle networks and introduce the main concept of Turbo CAN. Finally, its performance and backward compatibility are shown via analysis and computer simulations.

Prof. Ji-Woong Choi received the B.S., M.S., and Ph.D. degrees from Seoul National University (SNU), Seoul, South Korea, in 1998, 2000, and 2004, respectively, all in electrical engineering. From 2005 to 2007, he was a Postdoctoral Visiting Scholar at the Department of Electrical Engineering, Stanford University, Stanford, CA, USA. He also worked as a Consultant with GCT Semiconductor, San Jose, CA, USA, for the development of mobile TV receivers from 2006 to 2007. From 2007 to October 2010, he worked for Marvell Semiconductor, Santa Clara, CA, USA, as a Staff Systems Engineer for next-generation wireless communication systems, including WiMAX and LTE. Since October 2010, he has been with the Department of Information and Communication Engineering, Daegu Gyeongbuk Institute of Science and Technology (DGIST), Daegu, South Korea, as an Associate Professor. His research interests include communication theory, signal processing, vehicular communications and security, biomedical communication applications, and brain–machine interface. Prof. Choi received DGSIT Research Award in 2017, Postdoctoral Scholarship from Ministry of Information and Communication, Korea, in 2005, and Silver Award at Samsung Humantech Paper Contest in 2005.


Invited Talk 2: ”NTU-NXP Smart Mobility Test-Bed: A Campus-Wide Infrastructure for Connected Cars”
Dr. Yong Liang Guan, NTU Singapore, Singapore

V2X (vehicle to everything) communication refers to a new vehicular WiFi technology that allows moving cars to communicate not just directly with each other, but also with “access points” installed on lamp poles or roadside infrastructure. This technology promises to enhance road safety, cut driving time, save fuel, augment GPS, drive big data, and enable new road pricing. International standards have been defined. Market products have emerged. In this talk, I will give an overview of a campus-wide V2X test-bed jointly developed by NTU and NXP that conforms to the IEEE WAVE standard suite, the full-stack applications that the test-bed is capable of supporting, the V2X standardization landscape, and discuss some research projects related to this program.

Dr. Yong Liang GUAN (http://www.ntu.edu.sg/home/eylguan/) obtained his PhD degree from the Imperial College of London, UK, and Bachelor of Engineering degree with first class honors from the National University of Singapore. He is a tenured associate professor at the School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, where he was a Head of Division in 2011-2014 and the Director of the Positioning and Wireless Technology Center in 2007-2011. His research interests broadly include coding and signal processing for communication systems and data storage systems. He has published an invited monograph, 3 book chapters, and over 360 journal and conference papers. He is an Associate Editor of the IEEE Transactions on Vehicular Technology, and was an AE of the IEEE Signal Processing Letter. He has led 12 past and present externally funded research projects on V2V/V2I communication, wireless communication signal processing, coding for 10Tb/sq-in magnetic recording, acoustic telemetry for drilling application, etc., with total funding of over SGD 17 million. He has 3 granted patents.


Invited Talk 3: ”V2X Channel Characteristics: Guidelines for Designing V2X Service Applications”
Prof. Dong Seog Han, KNU, Korea

Vehicle-to-X (V2X) communication is no longer future oriented technology because the USDOT NHTSA recently issued a Notice of Proposed Rulemaking (NPRM) to mandate vehicle-to-vehicle (V2V) communication technology for new light vehicles in the United States. Establishing V2X application services needs special attention since vehicular communication environments are represented by low antenna heights and high mobility of vehicles. Such characteristics make the vehicular channel propagation dynamics more complex and time-varying. Thus, extensive analysis of V2X channel characteristics is essential for designing service applications. In this talk, both the state-of-the-arts and open issues on V2X channel characteristics including field measurement campaign and data analysis will be introduced.

Dong Seog Han received the B.S. degree in electronics engineering from Kyungpook National University, Daegu, Korea, in 1987, and the M.S. and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejon, Korea, in 1989 and 1993, respectively. From Oct. 1987 to Aug. 1996, he was with Samsung Electronics. Since Sep.1996, he has been with the School of Electronics Engineering, Kyungpook National University as a Professor. He worked as a courtesy Associate Professor in the Department of Electrical and Computer Engineering, University of Florida, in 2004. He was the director at the center of Digital TV and Broadcasting in the Institute for Information Technology Advancement (IITA) from July 2006 to July 2008. His main research interests are machine learning and vehicular communication systems.

Special Session III: IoT

08:20-10:00, October 19 (Thursday), 2017

Time Title Invited Speakers
08:20-10:00 A More Intuitive Way to Control IoT Devices in Proximity Prof. Kyunghan Lee,
UNIST, Korea
Calibrating Time-variant, Device-specific Phase Noise for COTS WiFi Devices Prof. Sangtae Ha,
University of Colorado, USA
Enabling Internet of Things: Observation, Challenges and Sensing Communication Technology Prof. D. Singh,
HUFS, Korea

Invited Talk 1: ”A More Intuitive Way to Control IoT Devices in Proximity”
Prof. Kyunghan Lee, UNIST, Korea

Controlling IoT devices is typically considered easy, but there are situations where this conception is wrong. Those include the followings: 1) when initializing an IoT device, 2) when trying to control an IoT device which is initialized and registered by another person, and 3) when trying to control IoT devices of the same type. We identify the challenges involved in these situations and propose a new intuitive communication method. Our method lets a user device such as a smartphone pinpoint and activate an IoT device with the help of an IR transmitter and communicate with the pinpointed IoT device through the broadcast channel of WiFi. Thanks to its unique design, our method allows a user device to immediately give a command to a specific IoT device in proximity even when the IoT device is uninitialized, unregistered to the control interface of the user, or registered but physically undistinguishable from others. Our method implemented on Raspberry Pi 2 devices demonstrate that it indeed enables intuitive and responsive controlling of IoT devices. Our experimental study shows that its end-to-end delay for each successful commanding to an IoT device is upper bounded by 2.5 seconds and has its median at about 0.74 seconds.

Kyunghan Lee received his B.S., M.S., and Ph.D. degrees in Electrical Engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2002, 2004, and 2009, respectively. He is currently an associate professor in the school of electrical and computer engineering at UNIST (Ulsan National Institute of Science and Technology), Ulsan, Korea. Prior to joining UNIST in 2012, he was with the Department of Computer Science, North Carolina State University, Raleigh, USA as a senior research scientist. His research interests include low latency networking, context-aware mobile computing, and human behavior modeling. Dr. Lee has received IEEE ComSoc William R. Bennett Prize twice (2013 and 2016) for his work on human mobility modeling and mobile data offloading. He has served as a technical program vice chair of IEEE SECON 2015 and a publication chair of IEEE MDM 2017. He also has served as a technical program committee of renowned conferences such as IEEE INFOCOM and ACM MobiSys.


Invited Talk 2: ”Calibrating Time-variant, Device-specific Phase Noise for COTS WiFi Devices”
Prof. Sangtae Ha, University of Colorado, USA

Current COTS WiFi based work on wireless motion sensing extracts human movements such as keystroking and hand motion mainly from amplitude training to classify different types of motions with training, as obtaining meaningful phase values is very challenging due to time-varying phase noises occurred with the movement. However, the methods based only on amplitude training are not very practical since their accuracy is not environment and location independent. This talk introduces a list of recent works on COTS WiFi based motion sensing and presents an effective phase noise calibration technique which can be broadly applicable to any COTS WiFi based sensing. We leverage the fact that multi-path for indoor environment contains certain static paths, such as reflections from wall or static furniture as well as dynamic paths due to human hand and arm movements. Our evaluation shows that calibrated phase values provide much rich, yet robust information on motion tracking — 70th percentile angle estimation error up to ten degrees, 80th percentile tracking error up to 15 cm, and its robustness to the environment and the speed of movement.

Sangtae Ha is an Assistant Professor in the Department of Computer Science at the University of Colorado Boulder. He received his Ph.D. in Computer Science from North Carolina State University. His research focuses on building and deploying practical network systems. He co-founded the Princeton EDGE Lab as its first Associate Director in 2009 and led its research team as an Associate Research Scholar at Princeton University from 2010 to 2013. He is a co-founder and the founding CTO/VP Engineering of DataMi, a startup company on mobile networks. He also co-founded Zoomi, which develops an adaptive learning platform using artificial intelligence. He is an IEEE Senior Member and serves as an Associate Editor for IEEE Internet of Things (IoT) Journal. He received the INFORMS ISS Design Science Award in 2014.


Invited Talk 3: ”Enabling Internet of Things: Observation, Challenges and Sensing Communication Technology”
Prof. D. Singh, HUFS, Korea

Sensing communication is a new concept of computing technology which is fast emerging as a successful extension to existing Internet in an embedded devices. Researchers have visualized interconnections of billions of smart embedded devices to change the way of life. There are several Internet-of-Things (IoT) and M-2-M initiatives going on to the development of the sensing technologies for smart city services especially in machine-to-real-world and machine-to-humans. The resultant of the sensing communication objects are to utilized embedded technologies to monitor, control for the comfortable and secure human life. In this talks, I would like to introduce a proposed distributed internet architecture model based on sensing communication services, which is based on hierarchical peer-to-peer networks an embedded 6LoWPAN devices and physical interconnections between objects and attributes to access the location. Finally, I will present real-time test-bed and simulation scenarios for the IoT applications such as women safety services, connected vehicles, global healthcare monitoring and military services especially for intelligent security management system.

Prof. (Dr.) Dhananjay Singh is the Director of ReSENSE Laboratory, an Assistant Professor with the Department of Electronics Engineering, and the Chairperson of the Division of Global IT, Hankuk University of Foreign Studies, Yongin, South Korea, since 2012. He has published 100+ refereed scientific papers, served 100+ TPC membership, and delivered 50+ invited talks into the major IEEE conferences/workshop. His research interests focus on the design, analysis, and implementation of algorithms/protocols for large-scale data set to solve real-world problems spec. Future Internet Architecture for Smart City and IoT services. He is a Senior Member of the ACM and IEEE Society. He was a Post-Doctoral Researcher and a Senior Member of Engineering Staff of Future Internet Architecture with the National Institute of Mathematical Sciences, and Electronics and Telecommunication Research Institute, Daejeon, South Korea, from 2010 to 2012. He has won Five times best paper awards from the IEEE conferences and two times fellowship award from APAN meeting for Singapore and Manila, Philippines. He received the B.Tech. degree in computer science and engineering from Veer Bahadur Singh Purvanchal University, Jaunpur, India, in 2003, the M. Tech. degree in wireless communication and computing from the Indian Institute of Information Technology, Allahabad, India, in 2006, and the Ph.D. degree in ubiquitous IT from Dongseo University, Busan, South Korea, in 2010.

Special Session IV: 5G Models and Technologies

16:00-17:40, October 19 (Thursday), 2017

Time Title Invited Speakers
16:00-17:40 5G Technology Overview Dr. Taeyoung Kim,
Samsung Electronics, Korea
5G – Massive, Ultra-Dense, and Cloud Computing Prof. Tony Quek,
SUTD, Singapore
Programmable and Agile Wired Infrastructure for 5G Dr. Sueng Yong Park,
Kulcloud, Korea

Invited Talk 1: ”5G Technology Overview”
Dr. Taeyoung Kim, Samsung Electronics, Korea

With the ever-increasing demands on mobile data traffic, stronger requirements on latency and reliability of novel mobile services, it becomes more challenging to meet those diverse needs by sheer migration of the existing cellular technologies. Consequently, 3GPP set out 5G standardization to accommodate those needs with evolutionary technologies in April 2016. This presentation introduces the candidate technologies to meet these diverse requirements and the progress of 3GPP standardization for 5G.
Samsung’s innovative and pioneering technologies are now being used in the most advanced 5G trials yet and are helping to lay the foundation of critical 5G standards. Through research partnerships spanning the industry value chain, Samsung and its global partners are now demonstrating the viability of some of the first practical use cases, including fixed wireless access.

Taeyoung Kim received the B.Sc., M.Sc., and Ph.D. degrees in electrical engineering from Yonsei University, Seoul, Korea, in 1998, 2000, and 2006, respectively. Since 2005, he has been with Samsung Electronics Company, Ltd., Suwon, Korea. He has years of experience in standardization of Mobile WiMAX (including IEEE802.16e/m) and 3GPP LTE. His current fields of interest include research/development of millimeter-wave beamforming system as next generation mobile communication system and advanced PHY algorithms


Invited Talk 2: ”5G – Massive, Ultra-Dense, and Cloud Computing”
Prof. Tony Quek, SUTD, Singapore

By 2020, the 5G systems will emerge as the next generation mobile communication technologies, where new services, applications and devices will drive requirements on data rate, ubiquity of data services, latency, cost, and reliability and further drive data traffic growth. Indeed, 5G systems will serve an unprecedented number of devices, providing ubiquitous connectivity as well as innovative and rate-demanding services. It is forecast that by 2020 there will be more than 50 billion connected devices, including not only human communications, but also machine communications. To meet such demanding challenges in 5G, new technologies have been developed and investigated all around the world. In this talk, we will review some of these technologies and investigate from a fundamental perspective to understand how these technologies can complement each other. Furthermore, we will address the implications of massive antennas, ultra-dense deployment, and cloud computing in 5G systems.

Tony Q.S. Quek received the B.E. and M.E. degrees in Electrical and Electronics Engineering from Tokyo Institute of Technology, Tokyo, Japan, respectively. At Massachusetts Institute of Technology (MIT), Cambridge, MA, he earned the Ph.D. in Electrical Engineering and Computer Science. Currently, he is a tenured Associate Professor with the Singapore University of Technology and Design (SUTD). He also serves as the Associate Head of ISTD Pillar and the Deputy Director of SUTD-ZJU IDEA. His current research topics include heterogeneous networks, wireless security, big data processing, and IoT.
Dr. Quek has been actively involved in organizing and chairing sessions, and has served as a TPC member in a numerous international conferences. He is serving as the Workshop Chair for IEEE Globecom in 2017, Track Co-Chair for IEEE VTC Spring in 2018, and the Track Co-Chair for IEEE PIMRC in 2018. He is currently an elected member of the IEEE Signal Processing Society SPCOM Technical Committee. He was an Executive Editorial Committee Member of the IEEE Transactions on Wireless Communications, an Editor of the IEEE Transactions on Communications, and an Editor of the IEEE Wireless Communications Letters. He is a co-author of the book “Small Cell Networks: Deployment, PHY Techniques, and Resource Allocation” published by Cambridge University Press in 2013 and the book “Cloud Radio Access Networks: Principles, Technologies, and Applications” by Cambridge University Press in 2016.
Dr. Quek received the 2008 Philip Yeo Prize for Outstanding Achievement in Research, the IEEE Globecom 2010 Best Paper Award, the 2012 IEEE William R. Bennett Prize, the IEEE SPAWC 2013 Best Student Paper Award, the IEEE PES General Meeting 2015 Best Paper, the 2015 SUTD Outstanding Education Awards – Excellence in Research, the 2016 Thomson Reuters Highly Cited Researcher, and the 2016 IEEE Signal Processing Society Young Author Best Paper Award.


Invited Talk 3: ”Programmable and Agile Wired Infrastructure for 5G”
Dr. Sueng Yong Park, Kulcloud, Korea

It is getting clear that the fifth generation (5G) wireless systems would bring a profound impact on the architecture of the modern communication infrastructure. The communication industry traditionally led the innovation by introducing the fast and convenient wireless and wired communication, paving a way to the modern information society. But, in recent years, the pace of such innovation is rather slowed and the communities are asking themselves if we can do better.
Very interestingly, the computer industry, in recent years, has been actively adopting the virtualization, SDN, software driven infrastructure, cloud computing and storage and other innovative infrastructure technologies. With the enhanced infrastructure, the industry has not only creatively broken the traditional business models but made the new and innovative models and is expanding market very quickly.
Proposals and new architectural models for 5G wireless systems are arguably the first pan-communication industry efforts to change the communication infrastructure from the base, making them as agile and programmable as those of computer industry. The scope of these works are too vast to cover in a short time. In this speech, we focus on the work in progress in the programmable and agile wired infrastructure, and study the impact of such new infrastructure on the business models of telecommunication industry.

Dr. Sueng Yong Park is a CTO of KulCloud, the first Korean SDN start-up, which released the first Korean SDN controller, first SDN controller export(to USA), first Openflow 1.0, 1.3, 1.4, 1.5 supports among other pioneering works. He is currently working on the SDN-based 5G infra technology, multi-domain service and infra orchestration and network analytics. Previously he was the Principle Investigator of Korea for the prestigious international multi-domain orchestration project, Federation for FIRE (Future Internet Research and Experiment), 2012-2016, which was governed by European Union 7th Framework Programme (EU FP7). The honorable project was highly selective and invited only one or two prestigious consortia, total of 15 organizations from 9 countries, including Korea, England, France, Germany, Spain, Australia, etc. For 4 years, the members of Fed4FIRE project integrated various virtual and physical resources, formalized management standard, invented schemes to manage IDs and transference of authorization and etc. The members of project are now actively contributing, MANO (Management and Network Operation) Standards, 3GPP management standardization, and ITU.
Before joining the PI of Fed4FIRE, Dr. Park worked for Cisco systems, San Jose, CA, U.S.A., and Samsung Electronics, Suwon, Korea for 5 years and 2 years respectively. He was the core developer of highly successful Cisco 7600 and 12000 series routers, and Samsung Ubigate router. He earned his BS from Yonsei Univ. (Electronic Engineering), MS and Ph.D. from University of Illinois at Urbana-Champaign, U.S.A. (Electrical and Computer Engineering).

Special Session V: Artificial Intelligence & Machine Learning

08:20-10:00, October 20 (Friday), 2017

Time Title Invited Speakers
08:20-10:00 Gauging Variational Inference Prof. Jinwoo Shin,
KAIST, Korea
From Big Data to Precision Oncology using Interpretable Machine Learning Prof. Su-In Lee,
University of Washington, USA
Exobrain: Natural Language Understand and Question Answering SW Dr. Kyoungman Bae,
ETRI, Korea

Invited Talk 1: ”Gauging Variational Inference”
Prof. Jinwoo Shin, KAIST, Korea

Computing partition function is the most important statistical inference task arising in applications of Graphical Models (GM). Since it is computationally intractable, approximate methods have been used to resolve the issue in practice, where mean- field (MF) and belief propagation (BP) are arguably the most popular and successful approaches of a variational type. In this paper, we propose two new variational schemes, coined Gauged-MF (G-MF) and Gauged-BP (G-BP), improving MF and BP, respectively. Both provide lower bounds for the partition function by utilizing the so-called gauge transformation which modifies factors of GM while keeping the partition function invariant. Moreover, we prove that both G-MF and G-BP are exact for GMs with a single loop of a special structure, even though the bare MF and BP perform badly in this case. Our extensive experiments confirm that the newly proposed algorithms outperform and generalize MF and BP.

Jinwoo Shin is currently an associate professor at the School of Electrical Engineering at KAIST, Korea. His current major research interest is on algorithmic questions for machine learning and networking. He obtained the Ph.D. degree from MIT in 2010 with George M. Sprowls (Best MIT CS PhD Thesis) Award. After spending two years (2010-2012) at Algorithms & Randomness Center, Georgia Institute of Technology, one year (2012-2013) at Business Analytics and Mathematical Sciences Department, IBM T. J. Watson Research, he joined KAIST EE in Fall 2013. He received Best Publication Award from INFORMS Applied Probability Society 2013 and ACM SIGMETRICS Rising Star Award 2015, in addition to best/oral papers at SIGMETRICS, NIPS and MOBIHOC. He is currently an associate editor of IEEE/ACM Transactions on Networking, ACM Modeling and Performance Evaluation of Computing Systems, and has served TPCs at AAAI, INFOCOM, INFORMS, MOBIHOC, NIPS, SIGMETRICS, WIOPT.


Invited Talk 2: ”From Big Data to Precision Oncology using Interpretable Machine Learning”
Prof. Su-In Lee, University of Washington, USA

While targeting key drivers of tumor progression (e.g., BCR/ABL, HER2, and BRAFV600E) has had a major impact in oncology, most patients with advanced cancer continue to receive drugs that do not work in concert with their specific biology. This is exemplified by acute myeloid leukemia (AML), a disease for which treatments and cure rates (in the range of 20%) have remained stagnant. Effectively deploying an ever-expanding array of cancer therapeutics holds great promise for improving these rates but requires methods to identify how drugs will affect specific patients. Cancers that appear pathologically similar often respond differently to the same drug regimens.
In this talk, I will present our on-going project on building a machine learning model that takes available molecular information, reasons about the best possible treatment strategy, and explains its reasoning. The most important step necessary to realize this goal is to identify robust molecular markers from available data to predict the response to each of hundreds of chemotherapy drugs. However, due to the high-dimensionality (i.e., the number of variables is much greater than the number of samples) along with potential biological or experimental confounders, it is an open challenge to identify robust biomarkers that are replicated across different studies. I will present two distinct machine learning techniques to resolve these challenges. These methods learn the low-dimensional features that are likely to represent important molecular events in the disease process in an unsupervised fashion, based on molecular profiles from multiple populations of patients with specific cancer type. I will present two applications of these two methods – AML and ovarian cancer. When the first method was applied to AML data in collaboration with UW Hematology and UW’s Center for Cancer Innovation, a novel molecular marker for topoisomerase inhibitors, widely used chemotherapy drugs in AML treatment, was revealed. The other method applied to ovarian cancer data led to a potential molecular driver for tumor-associated stroma, in collaboration with UW Pathology and UW Genome Sciences. Our methods are general computational frameworks and can be applied to many other diseases.

Professor Su-In Lee is an Associate Professor of Computer Science & Engineering and Genome Sciences at the University of Washington. She received her Ph.D. degree in Electrical Engineering from Stanford University in 2009. Before joining the UW in 2010, she was a Visiting Assistant Professor in the Computational Biology Department at Carnegie Mellon University.
Her interest is in developing advanced machine learning algorithms to analyze high-throughput data to discover molecular mechanisms of diseases, identify therapeutic targets, and develop personalized treatment plans given an individual’s molecular profile. She has been named an American Cancer Society (ACS) Research Scholar and received the National Science Foundation (NSF) CAREER award.


Invited Talk 3: ”Exobrain: Natural Language Understand and Question Answering SW”
Dr. Kyoungman Bae, ETRI, Korea

Exobrain project consists of three core technologies: natural language understanding technology for analyzing the language in a text; knowledge learning technology for accumulating the result of language analysis as the knowledge; question answering technology that infers the correct answer when a user enters a question in natural language. In this talk, I will introduce the detailed technologies of Exobrain; 1) Natural language understanding is a Korean language processing technology for understanding the meaning of vocabulary, grammar, and context in the text described in natural language. 2) knowledge learning is a technology for learning and generating new knowledge from big data. 3) question answering is a technology to understand questions composed of several sentences and deduce correct answers from the structured/unstructured knowledge.

Kyoungman Bae is a researcher in the language intelligence research group in the ETRI (Electronics and Telecommunications Research Institute). He received the Ph.D. degree in computer engineering from the Dong-A University, Korea, in 2016. His research field is a question classification and retrieval in cQA service. He is currently conducting research on KBQA and paraphrasing in the Exobrain project.

Special Session VI: Challenges and Opportunities in UAV Networks and Applications

10:30-12:10, October 20 (Friday), 2017

Time Title Invited Speakers
10:30-12:10 UAVs and Their Potential in Delivering IoT Services Prof. Tarik Taleb,
Aalto University, Finland
Surprise AI — A Novel Approach for Self-Monitoring/Alerting Intelligent Drones (and Self-Driving Cars, too!) Dr. Bo Ryu,
EpiSys Science, USA
UAV Management System Based on IoT Platform and U2X Networks Dr. Jaeho Kim,
KETI, Korea

Invited Talk 1: ”UAVs and Their Potential in Delivering IoT Services”
Prof. Tarik Taleb, Aalto University, Finland

The usage of Unmanned Arial Vehicles (UAVs), simply known as drones, for mail delivery, rescue team management, or disaster recovery operations is gaining lots of attention. Along with the maturity of the technology and relevant regulations, a worldwide deployment of these UAVs is expected. Whilst these drones would be deployed for specific objectives (e.g., mail delivery), they can be simultaneously used for offering numerous Value-Added Services (VASs), particularly in the area of Internet of Things (IoT), when they are equipped with suitable and remotely controllable sensors, cameras, and actuators. Indeed, integrating and orchestrating the different segments of drones (i.e., each manufactured with specific hardware and used for a specific purpose) would yield a potential Unmanned Arial System (UAS) that could be used as an important data transport platform, on the fly and in parallel to the existing Internet system. Sharing the infrastructure of this UAS for the provisioning of different IoT services would lower both capital and operational expenses, would encourage innovation giving birth to a plethora of new IoT services that can be offered only from height, and would create a novel ecosystem with new stakeholders. The edification of this self-* multi-purpose UAS along with its VAS and its orchestration system comes with a number of scientific challenges, that this talk will tackle, introducing some potential solutions.

Prof. Tarik Taleb is an IEEE Communications Society (ComSoc) Distinguished Lecturer and a senior member of IEEE. He is currently Professor at the School of Electrical Engineering, Aalto University, Finland. He is the director of the MOSA!C Lab (http://mosaic-lab.org/). Prior to his current academic position, he was working as Senior Researcher and 3GPP Standards Expert at NEC Europe Ltd, Heidelberg, Germany. He was then leading the NEC Europe Labs Team working on R&D projects on carrier cloud platforms, an important vision of 5G systems. Before joining NEC and till Mar. 2009, he worked as assistant professor at the Graduate School of Information Sciences, Tohoku University, Japan, in a lab fully funded by KDDI. From Oct. 2005 till Mar. 2006, he worked as research fellow at the Intelligent Cosmos Research Institute, Sendai, Japan. He received his B. E degree in Information Engineering with distinction, M.Sc. and Ph.D. degrees in Information Sciences from Tohoku Univ., in 2001, 2003, and 2005, respectively.
Prof. Taleb’s research interests lie in the field of architectural enhancements to mobile core networks (particularly 3GPP’s), mobile cloud networking, network function virtualization, software defined networking, mobile multimedia streaming, inter-vehicular communications, and social media networking. Prof. Taleb has been also directly engaged in the development and standardization of the Evolved Packet System as a member of 3GPP’s System Architecture working group. Prof. Taleb is a member of the IEEE Communications Society Standardization Program Development Board. As an attempt to bridge the gap between academia and industry, Prof. Taleb founded the “IEEE Workshop on Telecommunications Standards: from Research to Standards”, a successful event that got awarded “best workshop award” by IEEE Communication Society (ComSoC). Based on the success of this workshop, Prof. Taleb has also founded and has been the steering committee chair of the IEEE Conf. on Standards for Communications and Networking (http://www.ieee-cscn.org/).
Prof. Taleb is the general chair of the 2019 edition of the IEEE Wireless Communications and Networking Conference (WCNC’19) to be held in Marrakech, Morocco. He is/was on the editorial board of the IEEE Transactions on Wireless Communications, IEEE Wireless Communications Magazine, IEEE Journal on Internet of Things, IEEE Transactions on Vehicular Technology, IEEE Communications Surveys & Tutorials, and a number of Wiley journals. Till Dec. 2016, he served as chair of the Wireless Communications Technical Committee, the largest in IEEE ComSoC. He also served as Vice Chair of the Satellite and Space Communications Technical Committee of IEEE ComSoc (2006 – 2010). He has been on the technical program committee of different IEEE conferences, including Globecom, ICC, and WCNC, and chaired some of their symposia.
Prof. Taleb is the (co)recipient of the 2017 IEEE Communications Society Fred W. Ellersick Prize (May 2017), the 2009 IEEE ComSoc Asia-Pacific Best Young Researcher award (Jun. 2009), the 2008 TELECOM System Technology Award from the Telecommunications Advancement Foundation (Mar. 2008), the 2007 Funai Foundation Science Promotion Award (Apr. 2007), the 2006 IEEE Computer Society Japan Chapter Young Author Award (Dec. 2006), the Niwa Yasujirou Memorial Award (Feb. 2005), and the Young Researcher’s Encouragement Award from the Japan chapter of the IEEE Vehicular Technology Society (VTS) (Oct. 2003). Some of Prof. Taleb’s research work have been also awarded best paper awards at prestigious conferences.


Invited Talk 2: ”Surprise AI — A Novel Approach for Self-Monitoring/Alerting Intelligent Drones (and Self-Driving Cars, too!)”
Dr. Bo Ryu, EpiSys Science, USA

The use of Artificial Intelligence and Machine Learning (AI&ML) for autonomous systems such as drones and self-driving cars is already an old news. While the AI&ML-based prediction and classification capability continues to impress the world, the uncomfortable truth is that it is not failure-proof. For emerging autonomous systems such as drones and self-driving cars, prediction/classification errors, however small, may lead to costly, and even fatal consequences. In this talk, we introduce “Surprise AI”, a powerful learning paradigm capable of monitoring Deep Learning models in operation during runtime. Surprise AI can detect input data that will likely lead to prediction/classification errors in real time. We demonstrate, via extensive testing with several DL models available in the public domain, that Surprise AI successfully detects, during runtime, whenever the DL model encounters an input data which contains feature(s) that deviate from the model with as high as 99% accuracy.

Dr. Bo Ryu has over 20 years of experience as Principal Investigator and Program Manager for high-risk, high-payoff Research and Development (R&D) projects (sponsored by DARPA, NSF, NASA, and other US defense agencies. In 2012, he founded EpiSys Science, Inc., a technology start-up focusing on developing disruptive products for defense customers based on the systems science paradigm combining autonomous systems (e.g., drone control software and hardware), cognitive systems (e.g., software defined radios with cognitive networking), and digital signal processing (e.g., cognitive radios). Prior to founding EpiSys Science. He has authored/co-authored 50+ publications, and holds twelve (12) U.S. patents issued. He received Ph.D. in Electrical Engineering from Columbia University. He is the founding President of TeK One (http://tekone.org) aimed at promoting technology-based Korean-American start-ups.


Invited Talk 3: ”UAV Management System Based on IoT Platform and U2X Networks”
Dr. Jaeho Kim, KETI, Korea

Recently, commercial Unmanned Aerial Vehicles (UAVs) called drones have received a lot of interest in the public and civil environments for their possibility as a potential enabler of promising services such as policing, peacekeeping, surveillance, product deliveries, aerial photography, and agriculture. However, in order to support those services smoothly, there is needed a stable and reliable UAV control and management system. Regarding this issue, common platforms and U2X (UAV-to-UAV, UAV-to-Infrastructure) networks are required to manage UAVs. This talk shall concentrate on the R&D activities for IoT-based UTM (Unmanned Aerial System Traffic Management) system and U2X networks such as cellular-based U2X and FANET(Flying Ad hoc Network). As a case study, this talk shall also introduce an implementation and demonstration for the IoT based drone management system.

Jaeho Kim is a managerial researcher in IoT Platform Research Center at the Korea Electronics Technology Institute (KETI), South Korea from 2000. He is leading the research team for IoT Platform and Future Internet in KETI. His expertise covers wireless sensor networks, medium access protocols, and Internet of Things platforms. He is now serving as IoT Convergence Service Project Group chair of TTA (Telecommunications Technology Association) and Device Working Group chair of Korea IoT Association. He received a Ph.D in the electrical & electronic engineering from the Yonsei University, South Korea. His research interests are in the areas of Internet of Things and UAV communication and networks.