Prof. Frank Dignum
Bio: Frank Dignum got his PhD at the VU in Amsterdam in 1989. Since 2019 he is Wallenberg distinguished chair in AI at Umeå University in Sweden. He also has an affiliation to Utrecht University still and since 2013 he is a honorary principal research fellow of the University of Melbourne. Since 2014 he is a EurAI fellow. He is well known for his work on norms and his theory of social agents is employed in social simulations to support policy making and e-coaching. He published 18 books and more than 300 papers. At this moment his H-index is 54.
Some current projects include a social simulation to investigate the consequences and possible scenarios of the government restrictions and measures to combat the COVID-19 virus and social simulations to support policy making for human sustainable cities.
Title of talk: Socially aware AI for a human sustainable world
Abstract of talk: The major crises in the world at this moment do not have a single optimal solution. The refugee crises, COVID-19 situation, sustainable world all involve multiple stakeholders with their own interests and vulnerabilities. At the moment, most initiatives trying to contribute to solving these crises concentrate on one aspect and try to optimize on one aspect, such as logistics, epidemics or energy. However, all these aspects are related and an optimal solution found for one aspect might be bad for another aspect. E.g. locking people down to prevent the spread of the virus might have such a large impact on the economic and mental welfare of a country that it should not be done for more than one month.
People are both social and habitual animals. I.e. they will change their preference based on what they sense other people’s preferences are. If many people start going out in contravention of a regulation others might also do this and then suddenly people might change their preference in order to reap the benefits. Thus, when looking for solutions to optimize the effects of regulations one should keep in mind that people will adapt to the changing situation and can change preferences based on experiences and circumstances.
In my presentation, I will argue that we should be using AI in its social context:
Bio: Serena Villata is a Research Scientist in Computer Science, HDR, at the CNRS in the I3S Laboratory. Her research is in the fields of Artificial Intelligence and Natural Language Processing, and more specifically, in the field of artificial argumentation. Her main research subject is the automatic extraction of arguments and their relationships from text. Serena is also working on the definition of automatic reasoning models for decision-making by intelligent machines. An application of her work is dedicated to the cyberbullying instance detection on social networks. Her results are applied to the health and legal fields. Serena was invited to give an “Early Career Spotlight Talk" at the 27th International Joint Conference on Artificial Intelligence IJCAI-2018. She is the author of more than 120 scientific publications, and she coordinated and participated to many EU projects and industrial collaborations.
Title of talk: Artificial Machines Arguing for And with People
Abstract of talk: Argumentation is important for handling conflicting beliefs, assumptions, opinions, goals, and many other mental attitudes. Argumentation pervades human intelligent behavior, and I believe that it is a mandatory element to conceive autonomous artificial machines that can exploit argumentation models and tools in the cognitive tasks they are required to carry out. In this talk, I will present how artificial argumentation can be used to help people in having a better understanding of political debates by automatically identifying potential fallacies, in decision making on health records and clinical trials by highlighting evidence and claims, and in fighting fake news spread by generating a counter-argumentation.
Prof. Ole-Christoffer Granmo
Bio: Granmo is the Founding Director of CAIR: Centre for Artificial Intelligence Research, University of Agder. He obtained his master’s degree in 1999 and the PhD degree in 2004, both from the University of Oslo, Norway. Dr. Granmo has authored in excess of 140 refereed papers within machine learning, encompassing learning automata, bandit algorithms, Bayesian reasoning, reinforcement learning, and computational linguistics. Dr. Granmo is also the inventor of the Tsetlin Machine and a co-founder of the NORA: Norwegian Artificial Intelligence Consortium. Apart from his academic endeavors, he co-founded the company Anzyz Technologies AS.
Title of talk: Recent advances in Tsetlin machines
Abstract of talk: Tsetlin machines (TMs) are a new approach to machine learning that is founded on the Tsetlin Automaton by M. L. Tsetlin, a pioneering solution to the multi-armed bandit problem. TMs leverage frequent pattern mining and resource allocation principles to extract common patterns in the data, rather than relying on minimizing output error, which is prone to overfitting. Unlike the intertwined nature of pattern representation in neural networks, a TM decomposes problems into self-contained patterns, represented as conjunctive clauses. The clause outputs, in turn, are combined into a classification decision through summation and thresholding, akin to a logistic regression function, however, with binary weights and a unit step output function. Recent research reports several distinct TM properties. The patterns that a TM builds seem to be interpretable, similar to the branches of a decision tree. The TM supports convolution, providing competitive accuracy, learning speed, and memory footprint in comparison with CNNs, SVMs, Random Forests, Gradient Boosting, BinaryConnect, Logistic Circuits and ResNet. The TM has also achieved promising results in text classification by using the clauses to capture textual patterns. Further, Regression TMs compare favorably with Regression Trees, Random Forest Regression, and Support Vector Regression. Recently, TM hardware has demonstrated up to three orders of magnitude reduced energy usage and faster learning, compared to neural networks alike. In this talk, I will give a detailed introduction to the TM and cover the most recent developments, including regression TMs, convolution TMs, multi-layer TMs, clause weighting, clause indexing, novelty detection, and energy-efficient TM hardware architectures.
Prof. Sokratis Katsikas
Bio: Sokratis Katsikas was born in Athens, Greece, in 1960. He received the Diploma in Electrical Engineering from the University of Patras, Patras, Greece in 1982, the Master of Science in Electrical & Computer Engineering degree from the University of Massachusetts at Amherst, Amherst, USA, in 1984 and the Ph.D. in Computer Engineering & Informatics from the University of Patras, Patras, Greece in 1987. In 2019 he has awarded a Doctorate Honoris Causa by the Dept. of Production and Management Engineering of the Democritus University of Thrace, Greece. He is the Rector of the Open University of Cyprus, Nicosia, Cyprus, and Professor with the Center for Cyber and Information Security, Department of Information Security and Communication Technology, Norwegian University of Science and Technology, Norway. His research interests lie in the areas of information and communication systems security and of estimation theory and its applications. His research activity over the past 30 years has resulted in the publication of 39 books; 35 book chapters; 86 journal publications (of which 9 invited); and 130 publications in conference proceedings (of which 30 invited). According to Googlescholar, his research work has been cited more than 3.000 times and his h-index is 29. He has participated in more than 60 funded national and international R&D projects in his areas of research interest. He is serving on the editorial board of several scientific journals, and has served on/chaired the technical programme committee of more than 600 international scientific conferences. He chairs the Steering Committee of the ESORICS Conferences.
Title of talk: Integrating IT with OT: Trends and Challenges
Abstract of talk: The integration of Information Technology with Operation Technology, being one of the underlying and enabling factors of the Industry 4.0 initiative, brings about expectations for unprecedented value creation opportunities in industry. Unfortunately, these do not come without a price; in this case the price to pay is the increased vulnerabilities, the increased threats and the increased attack surface that result when industrial systems originally designed with little or no cybersecurity in mind connect to the Internet. Consequently, the cybersecurity of the IIoT becomes of paramount importance. Research has started focusing on this area, as well as on the related areas of cyber-physical systems security and industrial network security, but a multitude of issues still remain to be addressed. In this talk, we review recent research results on securely integrating IT and OT, with an eye towards identifying trends on one hand and areas where research seems to lag behind on the other.
Prof. Ibrahim A. Hameed
Bio: received the BSc degree in Electronic Engineering and MSc degree in Control Engineering from the Menofia University, Menofia, Egypt, in 1998 and 2005, respectively. Received Ph.D. degree in Industrial Systems and Information Engineering from Korea University, Seoul, South Korea and PhD degree in Mechanical Engineering from Aarhus University, Aarhus, Denmark in 2010 and 2012, respectively. From March 2011 to December 2012, he has been working as an Assistant Professor at Department of Industrial Electronics and Control Engineering, Menofia University, Menofia, Egypt. From January 2013 to July 2015, he has been working as a postdoc at Department of Electronic Systems, Aalborg University, Denmark. Hameed is currently a Professor at Dept. of ICT and Natural Sciences, Faculty of Information Technology and Electrical Engineering, Norwegian University of Science and Technology (NTNU), Norway. Hameed is a deputy head of research and innovation within the same department. He is the head of the international master program in simulation and visualization. He is the founder and the chair of the AI and robotics interest group, Norwegian Computer Association (Den Norske Dataforening DND), Norway. Hameed is an IEEE senior member and the elected chair of the IEEE Computational Intelligence Society (CIS) Norway section. He is the founder and head of the social robotics lab in Ålesund (SRLÅ). Hameed is the author of more than 120 journal and conference articles. His current research interest includes Artificial Intelligence, Machine Learning, Optimization, and Robotics.
Title of talk: Adversarial Attacks and Defense Strategies in Machine Learning
Abstract of talk: Machine learning techniques were designed assuming that the training and testing data are generated from the same statistical distribution. However, when those models are deployed, the statistical assumption may be internationally or unintentionally violated, leading to dramatic degradation of the performance of ones of the top-performing models. Therefore, machine learning models are usually re-trained on a set of samples collected during the network operation to adapt to the expected changes in the underlaying data distribution. Within this scenario, an attacker may attempt to poison the training data by injecting carefully designed samples to eventually compromise the whole learning process and try to seriously fool the deployed machine learning models leading to various misbehaviors. In this presentation, various white-box, gray-box and black-box attacks and defense strategies will be presented covering application areas that ranges from face and object recognition, audio recognition, NLP and time series classification.