Our presentation is centered around the development of provable cooperation schemes for sensor-based motion planning and interaction control of autonomous systems. Our goal is to design sound interfaces of our provable techniques with higher-level machine-intelligence based decision making schemes with the purpose of minimizing explicit communication. We address a decentralized motion planning & control architecture for cooperative loading task using heterogeneous robots operating in a constraint workspace with static obstacles. The optimal loading configuration is selected considering the connectivity of the space and the distance between the robots. A motion control scheme for each agent is designed and implemented in order to autonomously guide each robot to the desired loading configuration with guaranteed obstacle avoidance and convergence properties. The performance of the proposed strategy is experimentally verified in a variety of loading scenarios. We continue with a novel distributed leader-follower architecture for Cooperative Manipulation of Multiple Underwater Vehicle Manipulator Systems (UVMS) under Lean Communication. The leading UVMS, which has knowledge of the desired trajectory, tries to achieve tracking behavior via an impedance control law, leading the overall formation towards the goal configuration while avoiding collisions with the obstacles. The following UVMSs estimate the object's desired trajectory via a novel prescribed performance estimation law and implement a similar impedance control law. No explicit data is exchanged online among the robots. Various simulations and experiments clarify the proposed method and verify its efficiency.

Dr. Kostas J. Kyriakopoulos was born in Athens, Greece in 1962 and received a Diploma in Mechanical Eng (Honors) from the National Technical University of Athens (NTUA), in 1985 and the MS & Ph.D. in Computer & Systems Engineering from Rensselaer - ECSE, Troy, NY in 1987 and 1991, respectively. From 1988 to 1991 he did research at the NASA Center for Intelligent Robotic Systems for Space Exploration. Between 1991-93 he was an Assistant Prof. at ECSE - RPI and the NY State CAT in Automation and Robotics. Since 1994 he has been with the Mechanical Engineering Department at NTUA, Greece, where he serves as a Professor, Director of the Post-Graduate Program on "Automation Systems", Director of the Control Systems Lab and Director of the Departmental Computation Lab. His interests are in Nonlinear Control and Embedded Systems applications in Sensor Based Motion Planning & Control of multi-Robotic Systems: Manipulators & Vehicles (Mobile, Marine and Aerial). He was awarded the G.Samaras award of academic excellence (NTUA), the Bodosakis Foundation Fellowship (1986-1989), the Alexander Onassis Foundation Fellowship (1989-1990) and an Alexander Von Humboldt Fellowship (1993). Dr. Kyriakopoulos has published more than 320 papers in journals and refereed conferences; he is Specialty Chief Editor for "Frontiers in Robotics and AI" and he serves in the editorial committees of a number of journals and conferences, while he has served as an administrative member of a number of international conferences. He has acted as a PI in 35 R & D projects, half of which were funded by the European Commission. He is an IEEE Fellow.

The Encoder-Decoder models based on Recurrent Neural Network (or Seq2Seq model in brief) or its variant such as Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM) cell has demonstrated decent success in many areas. However, it is still a mystery why a recurrent network with such simple structure can achieve various complex tasks. In this talk, I will first demonstrate the power of a Seq2Seq model in learning hidden connection between inputs and outputs, and then offer an in-depth analysis to the neural level to explain how and why it works. Finally I will describe a general strategy to perform in-depth analysis for deep recurrent neural networks.

Shou-de Lin is currently a full professor in the CSIE department of National Taiwan University. He holds a BS degree in EE department from National Taiwan University, an MS-EE degree from the University of Michigan, an MS degree in Computational Linguistics and PhD in Computer Science both from the University of Southern California. He leads the Machine Discovery and Social Network Mining Lab in NTU. Before joining NTU, he was a post-doctoral research fellow at the Los Alamos National Lab. Prof. Lin's research includes the areas of machine learning and data mining, social network analysis, and natural language processing. His international recognition includes the best paper award in IEEE Web Intelligent conference 2003, Google Research Award in 2007, Microsoft research award in 2008, 2015, 2016 merit paper award in TAAI 2010, 2014, 2016, best paper award in ASONAM 2011, US Aerospace AFOSR/AOARD research award winner for 5 years. He is the all-time winners in ACM KDD Cup, leading or co-leading the NTU team to win 5 championships. He also leads a team to win WSDM Cup 2016. He has served as the senior PC for SIGKDD and area chair for ACL. He is currently the associate editor for International Journal on Social Network Mining, Journal of Information Science and Engineering, and International Journal of Computational Linguistics and Chinese Language Processing. He is also a freelance writer for Scientific American.

In this talk, I will explain how human evolution has shaped the way we respond to and interact with media and information technologies in certain ways. After that, I will explain how our understanding of media psychology can be effectively applied to the design of successful ICT (information and communication technologies) innovation. I propose that in order to design innovative ICT products and services, we need to understand and develop the following four things--1. Evolutionary responses to technology form factors; 2. Sensor technologies to understand users; 3. Technologies manifesting machine's wills and desires; 4. Social interfaces. Examples come from both my academic and industry works spanning from evolutionary psychology to current directions in smart media design by Samsung and other ICT companies.

Kwan Min LEE (Ph.D., Stanford; MA, MSU; BA, Sogang) is the inaugural Korea Foundation Professor in Contemporary Korean Society and New Media, Fellow of Nanyang Technopreneurship Center (NTC), and Director of UX (User Experience) Lab at the Wee Kim Wee School of Communication and Information at Nanyang Technological University (NTU). Previously, Lee was the founding director of Interaction Science Research Center and the founding WCU (World Class University) Professor of the Department of Interaction Science at Sungkyunkwan University (SKKU), S. Korea. Lee also directed Samsung Electronics' User Experience (UX) Group and the Creative Lab (C-Lab) as one of the youngest Vice Presidents in the Samsung corporate history. Prior to SKKU and Samsung, Lee had taught at the Annenberg School for Communication and Journalism at the University of Southern California (USC) for 12 years as one of the youngest tenured professors. Lee specializes in UX (User Experience) research and design, social and psychological effects of ICT (Information and Communication Technologies), and human-machine interaction. Lee's works have been widely adopted as major course readings at leading academic institutes such as Stanford, Penn, and MIT. His research findings have been covered by Washington Post, BBC News, USA Today, and other major news agencies. Lee is an elected Fellow of the International Communication Association (ICA), one of the highest academic honors in the field of Communication. Lee's research funding exceeds 12 million US dollars. His research has been supported by the National Science Foundation (USA), the Ministry of Education, Science, and Technology (S. Korea), Ministry of Education (Singapore), Samsung Electronics, and Hyundai Motors. Lee holds multiple international (USA, EU, and S. Korea) patents in smart display interfaces, remote controller, gesture control, and multimedia production. He has been elected as Fellow of International Communication Association (ICA), one of the highest academic honors in the field of Communication, in 2019.

Deep learning has achieved great success in various tasks, when it is trained with large amount of labeled data. However, it is a challenging task to train deep models using only a handful of labeled examples. Few-shot learning, the goal of which is to learn from only a few examples in each class label, has been a popular subject in machine learning and computer vision communities. Learning to learn, which is also known as meta-learning, has emerged again as a promising approach to tackle few-shot problems. In this talk, I introduce recent advances in meta-learning, the underlying idea of which is to leverage past experience to learn a prior over tasks, so that it can quickly adapt to a novel task. I begin with a few metric-based deep learning methods that have been developed to solve a few-shot learning task. I give an information-theoretic unified view of existing metric learning methods and present a method for learning discrete latent representation shared across relevant tasks to enable a model to adapt to new tasks quickly. Experiments show that DIMCO requires less memory (i.e., code length) for performance similar to previous methods and that our method is particularly effective when the training dataset is small. I also illustrate gradient-based meta-learning, such as model-agnostic meta-learning (MAML), emphasizing my recent work on MT nets that learn layer-wise subspace and metric from a set of tasks. Finally I also introduce a method to warm-start Bayesian optimization which is a critical technique for AutoML, experimental design, or neural architecture search.

Seungjin Choi received B.S. and M.S. degrees in electrical engineering from Seoul National University, Korea,n in 1987 and 1989, respectively, and a Ph.D. degree in electrical engineering from the University of Notre Dame, Indiana, in 1996. He was with the Laboratory for Artificial Brain Systems, RIKEN, Japan, in 1997, working with Prof. Andrew Cichocki and Prof. Shunichi Amari on independent component analysis as a Frontier researcher. He was an Assistant Professor in the School of Electrical and Electronics Engineering, Chungbuk National University from 1997 to 2000. From 2001 to 2019, he was a Professor of Computer Science at Pohang University of Science and Technology, Korea. He was the director of Machine Learning Research Center where about 15 professors in top five universities participated. He has held consulting professor position for Shinhan Big Data Center, Samsung Research, Samsung Advanced Institute of Technology. He is currently a president of AI Society in KIISE. He is serving or has served as Area Chairs or Senior Program Committee for top-tier machine learning or AI conferences, including NeurIPS, ICML, AISTATS, AAAI, IJCAI. His primary research interests include statistical machine learning and probabilistic inference. Recently he is working on meta-learning and Bayesian optimization.

Nowadays, sensing technologies and large-scale computing infrastructures have produced a variety of big data in urban spaces, e.g. human mobility, air quality, traffic patterns, and geographical data. The big data implies rich knowledge about a city and can help tackle these challenges when used correctly. That is, holistic urban big data plays the key role in smart city constructions. However, processing urban big data needs the specific computing engine different with traditional one such as Hadoop and Spark, because the sensing and knowledge representation are more complicated than domain-specific big data. In this talk, we will give some properties for processing urban big data and introduce a new platform for processing and analyzing urban big data. Then we discuss how the collaborative computing bridges the data and computation in the cyber space and the environment, systems, people and things in the physical world.

Minyi Guo received the BSc and ME degrees in computer science from Nanjing University, China; and the PhD degree in computer science from the University of Tsukuba, Japan. He is currently Zhiyuan Chair professor of Shanghai Jiao Tong University (SJTU), China. Before joined SJTU, Dr. Guo had been a professor of the school of computer science and engineering, University of Aizu, Japan. Dr. Guo received the national science fund for distinguished young scholars from NSFC in 2007, and was supported by "Recruitment program of Global Experts" in 2010. His present research interests include parallel/distributed computing, compiler optimizations, embedded systems, pervasive computing, big data and cloud computing. He has more than 400 publications in major journals and international conferences in these areas. He received 7 best/highlight paper awards from international conferences including ALSPOS 2017 and ICCD 2018. He is now on the editorial board of IEEE Transactions on Parallel and Distributed Systems, IEEE Transactions on Cloud Computing and Journal of Parallel and Distributed Computing. Dr. Guo is a fellow of IEEE, and a fellow of CCF.

The Transformed Role of the Viewer: Second Screens and the Social Soundtrack

The nearly ubiquitous use of mobile devices integrated with easy interface with social media platforms facilitates a unique social interaction about broadcast media and other events that alters the role of the viewer from a passive to an active function, with the visitor engaged in information sharing, consumption, and dissemination often in real time. This technology affordance for online conversation about an event is referred to as the second screen phenomenon, although there may be multiple (i.e., more than two) screens involved. The resulting online conversation from second screen interaction about an event is referred to as the social soundtrack. The social soundtrack is an interesting conversational form of information sharing, information interaction, and information diffusion. This keynote will introduce the theoretical constructs and empirical measures of social soundtrack and second screen research, along with application of these constructs and measures in current investigations involving millions of posts on multiple social media platforms. Research concerning social soundtrack and secondary screens is important in identifying the influence and affordances that technology has on social media conversations from an information sharing. Research findings can also shed light on social communication in relationship to the cultural impact of broadcast media events, the social interaction in cross technology usage for second screens, and the effect of second screen technologies on pop culture and human information processing.

3D Reconstruction for Object Modeling and Scene Analysis

3D reconstruction is an important field in computer vision, and results accumulated in the field have found wide applications in virtual reality, creative media design, and robotics. But nevertheless, we still face great challenges when we try to use the techniques in modeling both objects with complex structures or large-scale scenes. The major difficulties come from several constraints in traditional approaches, including ambiguity and uncertainty inherent in the reconstruction algorithms, limitation on viewpoint movements, occlusion of objects, and low-resolutions of available 3D data. In the talk, I will introduce some newly developed methods aiming to solve the problems by making good use of imaging geometry principles and fusion of data from different sensors. Main topics include: reconstruction from silhouettes from a camera system with two planar mirrors; depth image super-resolution based on similarity-aware patchwork assembly; urban scene description by analysis of 3D data collected from car-mounted sensors. I also will report results from an application of such 3D digitization techniques in heritage documentation, mainly for grotto objects and scenes.

Intelligence in Education

Social software such as online forums, Wikis, and social networking sites, plays an important role in various fields, including science, politics, and education. Our goal is to analyze social activities within online communication and collaboration environments, and develop computational tools that support and promote effective interactions and participation. This talk present our work on online discussion modeling and intelligent tools for assisting discussion participants. We first analyze how messages and individual discussants contribute to Q&A discussions. We present a model for capturing information seeking or information providing roles of messages, such as question, answer or acknowledgement. We also identify user intent in the discussion as an information seeker or a provider. We show how the role information can be combined with linguistic and temporal features for developing a predictive model of discussant performance. We also demonstrate how such role information can be used for promoting interactions among potential peer collaborators.
In the latter part of the presentation, we show how such analyses can be a powerful tool for dialogue mediators and participants. In particular, we present a computational workflow (big data) framework that enables efficient and robust integration and analyses of diverse datasets. The analysis results are used for assisting discussion mediators or facilitating just-in-time adaptation to discussants' needs, such as identifying unresolved issues or help seekers who need more assistance.

The iSchool Vision for Interdisciplinary University Research and Education

The emergence of iSchools has been much discussed, with respect to an interdisciplinary vision for both research and education of undergraduate and graduate students. The Pennsylvania State University was one the first iSchools, launched in 1998 to meet the needs for workforce development of students who have the skills of information technology but also to research topics in real world interdisciplinary computing. In this talk, I will give a brief history of how and why this new realm of academic pursuits has emerged, illustrated throughout with examples drawn from education and research activities at Penn State and other iSchools. Reflecting on the past 15 years, I will also point to a set of continuing challenges and opportunities for interdisciplinary study that is founded on the integration of the information sciences, an increasingly ubiquitous technological substrate, and the broad and ambiguous implications of human individuals and organizations situated in real world activities.

Wireless HD Video Transmission Technology: Challenges and Future Applications

Bio-Inspired Supercomputing and Big Data

In order to convert data to knowledge, it is necessary to search (+process) data sets that are on the order of zettabytes in size (Big Data). Conventional computers (uniprocessor systems) are unable to process Big Data in a timely manner. Inherent limitations on the computational power of sequential uniprocessor systems have lead to the development of parallel multiprocessor systems. The two major issues in the formulation and design of parallel multiprocessor systems are algorithm design and architecture design. The parallel multiprocessor systems should be so designed so as to facilitate the design and implementation of the efficient parallel algorithms that exploit optimally the capabilities of the system. From an architectural point of view, the system should have low hardware complexity, be capable of being built of components that can be easily replicated, should exhibit desirable cost-performance characteristics, be cost effective and exhibit good scalability in terms of hardware complexity and cost with increasing problem size. In distributed memory multiprocessor systems, the processing elements can be considered to be nodes that are connected together via an interconnection network... The design presented in this talk is bio-inspired.

Renewable Energy Powered, Disaster-Resilient Wireless Network Infrastructure

The Positioning of Non-Volatile Memory in Embedded System Designs

Cyber-Physical and Networked Sensor Systems: Challenges and Opportunities

Beyond Ubiquitos Computing: The HoneyBee Ensemble Computing Environment

Living Interactively and Socializing Ubiquitously

Next Generation Network, Wireless Network and Topology Control with Small Routing Cost