The review of the TEWI colloquia of Prof. Wen-Huang Cheng from May 10, 2017 comprises the slides.

Slides:

Abstract and Bio:

People are interested in predicting the future. For example, which films will bomb or who will win the upcoming Grammy awards? Making predictions about the future is not only fun matters but can bring real value to those who correctly predict the course of world events, such as which stocks are the best purchases for short-term gains. Predictive analytics is thus a field that has attracted major attention in both academia and the industry. As social media has become an inseparable part of modern life, there has been increasing interest in research of leveraging and exploiting social media as an information source for inferring rich social facts and knowledge. In this talk, we will address an interesting and challenging problem in social media research, i.e., predicting social media popularity. We aim to discover which image posts on social media are the “stars of tomorrow”, those will be the most engaging for social media audiences, e.g., receiving the most likes. Sociological finding and our novel solutions to effectively develop a structured modeling for popularity dynamics will be presented.

Wen-Huang Cheng received the B.S. and M.S. degrees in computer science and information engineering from National Taiwan University, Taipei, Taiwan, in 2002 and 2004, respectively, where he received the Ph.D. (Hons.) degree from the Graduate Institute of Networking and Multimedia in 2008. He is an Associate Research Fellow (Associate Professor) with the Research Center for Information Technology Innovation (CITI), Academia Sinica, Taipei, Taiwan, where he is the Founding Leader with the Multimedia Computing Laboratory (MCLab), CITI, and an Associate Research Fellow with a joint appointment in the Institute of Information Science. Before joining Academia Sinica, he was a Principal Researcher with MagicLabs, HTC Corporation, Taoyuan, Taiwan, from 2009 to 2010. His current research interests include multimedia content analysis, multimedia big data, deep learning, computer vision, mobile multimedia computing, social media, and human computer interaction. He has received numerous research awards, including the 2016 Y. Z. Hsu Scientific Paper Award, the 2015-2016 Presidential Achievement Award of Rotary International, the Outstanding Youth Electrical Engineer Award from the Chinese Institute of Electrical Engineering in 2015, the Top 10% Paper Award from the 2015 IEEE International Workshop on Multimedia Signal Processing, the Outstanding Reviewer Award from the 2015 ACM International Conference on Internet Multimedia Computing and Service, the Prize Award of Multimedia Grand Challenge from the 2014 ACM Multimedia Conference, the K. T. Li Young Researcher Award from the ACM Taipei/Taiwan Chapter in 2014, the Outstanding Young Scholar Awards from the Ministry of Science and Technology in 2014 and 2012, the Outstanding Social Youth of Taipei Municipal in 2014, the Best Reviewer Award from the 2013 Pacific-Rim Conference on Multimedia, and the Best Poster Paper Award from the 2012 International Conference on 3D Systems and Applications. He is APSIPA Distinguished Lecturer.

The review of the TEWI colloquia of Dr. Dimitrios Serpanos from March 9, 2017 comprises the video recording and slides.

Video:

No correct api key set in Bitmovin Settings.

Slides:

Abstract and Bio:

Industrial control systems (ICS) control and manage a significant portion of critical infrastructure. As cyberattacks increasingly target critical infrastructure, ICS security and resilience are required to avoid catastrophic events that may lead even to loss of life. Importantly, ICS differ from traditional IT systems in several ways, from interacting with physical processes to requirements for continuous operation and real-time processing.

In this talk, we present a behavior-based approach to the design of secure and resilient industrial control systems. Starting with a programmable specification of a control process, we develop executable code with specified security properties. A run-time middleware, ARMET, monitors the execution of the program, identifying behavioral deviations due to intrusions or process failures and leading to diagnosis and system recovery. Importantly, our approach includes a novel method for vulnerability analysis of processes.

Dimitrios Serpanos is Professor of Electrical and Computer Engineering, University of Patras and Director of the Industrial Systems Institute in Greece. He has worked at IBM Research, on the faculties of the University of Crete and the University of Patras, as Principal Scientist at QCRI, as President of the University of Western Greece, and has served another term as Director of the Industrial Systems Institute. He holds a Ph.D. in Computer Science from Princeton University. His research is in embedded systems, industrial systems and security.
Professor Serpanos has co-authored several books and published research work extensively. His research has been funded by both the European Commission, the Greek Government and industry in the EU and the USA.

Montag, 10. April 2017, E.1.42

8:30 – 9:30, Christoph BENZMÜLLER

Title: Calculemus! Progress in Universal Logic Reasoning and Computational Metaphysics

Abstract: Inspired by Leibniz‘ famous vision „Calculemus! — Let us calculate“, an approach to universal logic reasoning is presented. The proposed solution is indirect: it utilises classical higher-order logic as a universal meta-logic in which various other (classical and non-classical) logics from today’s logic zoo can be shallowly embedded and combined; the embedded (object-)logics can then be employed for the representation of knowledge in different application domains.

This semantical embedding approach suppor ts the formalisation and deep logical assessment of rational arguments on the computer. In this sense, it partially realises Leibniz‘ dream. To illustrate this, I will present a rigorous analysis of modern variants of the “Ontological Argument for the Existence of God“, including Kurt Gödel’s seminal contribution. By utilising the approach, even relevant new insights about the ontological argument have recently been revealed by automated theorem provers. These research activities have inspired the conception of a new, awarded lecture course on “Computational Metaphysics“ at Freie Universität Berlin, which brings together students from computer science, mathematics and philosophy.

The approach, which fruitfully combines and exploits, amongst others, own research contributions from the last two decades, is by no means  restricted to metaphysics or philosophy; this is evidenced by recent applications e.g. in mathematics and artificial intelligence.

11:00 – 12:00, Wolfgang FABER

Titel: Deklarative Formalismen als Motoren Semantischer Systeme

Abstract: Semantik in Wissensbasierten System ist der Schlüssel, um über das vorhandene Wissen selbst Schlüsse zu ziehen. Semantik kommt besonderer Bedeutung zu, wenn nicht nur Antworten oder Lösungen gefunden werden sollen, sondern auch Begründungen oder Erklärungen selbiger. In diesem Vortrag soll die Rolle deklarativer Formalismen (im Sinne deklarativer Sprachen und Systeme) in solchen Semantischen Systemen herausgearbeitet werden. Besondere Berücksichtigung werden dabei logische Programmierung und Regelsysteme finden.

14:30 – 15:30, Martin GEBSER

Titel: Knowledge Representation and Reasoning in Practice

Abstract: Modern computing systems provide tremendous resources to automate sophisticated calculations, as required in various application areas, including life sciences, machine learning, planning and scheduling, product configuration, system synthesis and verification, etc. This goes along with an increasing demand for expressive, general-purpose means for knowledge representation and reasoning. In my talk, I outline how computational logic along with extensions can serve as a powerful paradigm for representing and solving demanding real-world applications by means of general, domain-independent search and optimization methods.

Dienstag, 11. April 2017, E.1.42

8:30 – 9:30, Jörg HOFFMANN

Titel: Automatic Planning: Some Recent Methods and Applications

Abstract: Automatic Planning is one of the founding areas of Artificial Intelligence, concerned with the automation of problem solving: the design of algorithms whose input is a declarative problem description („the rules of the game“) and that automatically find a solution („the game strategy“) to any such input. The curse of generality is complexity — even the simplest planning problems are PSPACE-complete to solve, in the size of the declarative description — so effective search methods are paramount. After a brief introduction to the area, the talk summarizes two such methods developed in recent work by the author, namely star-topology decomposition which breaks conditional dependencies among components in the input; and conflict-based learning which learns sound and generalizable knowledge from dead-end states encountered during forward state space search. We then highlight a recent application of planning techniques, to simulated network penetration testing in IT security. We conclude with an outlook on future challenges.

11:00 – 12:00, Stefan EDELKAMP

Title: Multi-Goal Motion Planning

Abstract: In the clothes of vehicle routing, multi-goal task planning has been an apparent optimization question for a fairly long time. We will present advances in high-speed solving these discrete planning problems. With fast single-source shortest path planning, the in- or outdoor map for the vehicle fleet is condensed to a graph of customer orders, which could be any combination of pickups and deliveries.  For one vehicle this setting corresponds to a variant of the NP-hard traveling salesman problem with capacities and time windows, while for several vehicles and according to the dynamics in the orders within a multi-agent simulation system, we support negotiating agents that offer multi-modal transport services.

While the discrete multi-goal task planning problem is already hard, the integration of task and motion planning is considered to be one of the most important challenges to nowadays robotics. Robots have sizes, heading, and velocity, and their motion can often be described only according to non-linear differential equations. The dynamics of movements, existing obstacles and many waypoints to visit are only some of the challenges to face. In real-world problems, we often have additional constraints like inspecting areas of interest in some certain order, while still minimizing the time for the travel. The trickiest part is to solve the hard combinatorial discrete tasks like the generalized and clustered traveling salesman problems, and -at the same time- providing valid trajectories for the robot. We use a framework in which a motion tree is steadily grown, and abstractions to discrete planning problems are used as a heuristic guidance for the on-going solution process to eventually visit all waypoints. In case of inspection, we generate the waypoints fully automatically, using a combination of 2D and 3D skeletonization together with a filtering mechanism based on hitting sets.