CETINIA Invited lectures:

Marc Serramia Amoros, City - University of London (UK)

• Title: Value-alignment and its applications in policy-making

• Date & Venue: 14/11/2024, 11:30h, URJC Móstoles, Salón de Grados 101, Departamental II

• Abstract:  With the advancement of AI, it has become paramount that we ensure it behaves and makes decisions in line with our moral values and priorities. This problem, called value-alignment, presents two main challenges: to formalise values mathematically, and to show how they can be used to regulate behaviour. In this seminar, we will see how we can address these challenges. First, we will define moral values in a computer-readable way based on the ethics literature. Second, we will explore the use of value-aligned norms to regulate agent behaviour. We will also show how these two contributions can easily be adapted for value-aligned decision making. Finally, we will illustrate how this theoretical research can be of practical use in policymaking, in particular for participatory budgets, processes where citizens propose and decide how to spend a government-defined budget. Value-alignment techniques can resolve issues like low participation, underrepresentation of minorities, and under-funding.

• Short Bio: Marc Serramia Amoros is a Lecturer in Computer Science at City, University of London and a member of CitAI. His research interests are Multiagent Systems and AI Ethics. In particular, Normative Multiagent Systems, Value alignment, Privacy, and applications of AI in policymaking (more specifically, participatory budgets). Previously, he was a postdoctoral researcher at King’s College London. His research there focused on finding ways to produce prescriptive norms to ensure smart assistants behave in a way aligned to what their users expect, for example, with regard to their privacy preferences. He obtained his PhD from the University of Barcelona and the Artificial Intelligence Research Institute of the Spanish National Research Council (IIIA-CSIC). His PhD thesis explored value alignment in normative multiagent systems, and received the best thesis award by the Catalan Association of Artificial Intelligence. He also received the Young Researchers award of the Spanish Scientific Society of Computer Science and the BBVA Foundation..

• Slides

Viviana Mascardi, University of Genoa (Italy)

• Title: Extended reality, chatbots and intelligent agents: synergies and perspectives

• Date & Venue: 18/10/2024, 11h, URJC Quintana (Calle Quintana 21,  Madrid), Room 504 (5th floor)

• Abstract:  The recent advent of the Metaverse, the market expansion of the game industry, the growing success of virtual and augmented reality in research, training, education, and rehabilitation, the need for realistic simulations of complex scenarios, all show that the non-physical reality where people may meet, play, learn, is becoming more and more important in our daily lives. This expansion calls for injecting autonomy, proactivity, reactivity, social ability - including conversational capability into the virtual characters living in the extended reality environment. This talk will explore the existing and the visionary synergies between extended reality and intelligent software agents.

• Short Bio: Viviana Mascardi is Associate Professor in Computer Science at the University of Genova (UniGE), Italy, DIBRIS (Dipartimento di Informatica, Bioingegneria, Robotica e Ingegneria dei Sistemi). Her research interests include modeling, runtime verification, rapid prototyping, and development of platforms for complex and distributed systems (multiagent systems, MASs), mainly based on computational logic and declarative agent languages and technologies; ontologies; interplay between MAS, ontologies, and Augmented/Virtual Reality; applications of agents and MASs. Viviana Mascardi is the chair of the European Association for Multi-Agent Systems (EURAMAS) since April 2022 and served as Program Co-Chair at AAMAS, the International Conference on Autonomous Agents and Multiagent Systems, in 2022. She had organizational roles in many Artificial Intelligence conferences including IJCAI and ECAI, besides AAMAS. 

• Slides

Franziska Klügl, University of Örebrö (Sweden)

• Title: Explicitness for Credibility of Agent-Based Simulations

• Date & Venue: 17/10/2024, 12:30h, URJC Móstoles, Salón de Grados 101, Departamental II

• Abstract:  Agent-based simulation can be seen as a well-established application of multi-agent systems. Yet, one still experiences hesitation in accepting agent-based approaches in decision support or in some scientific domains. Are there still challenges for credibility of models and simulation results? In the presentation, I will discuss those and talk about a number of approaches developed to enhance the explicitness of the model, exemplified by a economic model analyzing dynamics of technology uptake in manufacturing services.

• Short Bio: Franziska Klügl is Professor in Computer Science at the School of Science and Technology of örebrö University, Sweden. Her main research interests are related to languages, processes and tools for multi-agent systems, in particular when applied to simulation modeling. Her vision is to develop the approaches to agent-based simulation modelling so that domain experts without formal training in computer science can produce and use reliable, credible and valid models. One of the results of the research in Franziska's group is SeSAm - a visual programming tool for agent-based simulation that allows rapid prototyping as well as convenient development of complex agent-based simulation models.

• Slides

Giuseppe Vizzari, University of Milano-Bicocca (Italy)

• Title: AI and Simulation: a composite relationship

• Date & Venue: 17/10/2024, 11:15h, URJC Móstoles, Salón de Grados 101, Departamental II

• Abstract:  Results of AI are apparent and they triggered legitimate research questions on potential implications on many different fields. A recent Gartner hype cycle reports the topic “AI Simulation” in the "innovation trigger” phase, with a growing expectation for results in a 2 to 5 years timeframe. The relationship between simulation and AI, however, is not new: AI has already been used in different phases of a simulation project, sometimes within simulation models, but also for analysis of data describing real world dynamics, or simulation results. This talk will discuss a general framework of simulation projects, presenting concrete cases in which the paths of simulation and AI have already crossed, trying to conclude with some lessons learned and lines for future developments.

• Short Bio: Giuseppe Vizzari is Associate Professor at the Department of Informatics, Systems and Communication of University of Milano-Bicocca, Italy. His research activities mainly concern agent based models and technologies (situated agent models and applications, environments for multi-agent systems, agent-based modelling and simulation of complex systems, with particular attention to crowds of pedestrians and traffic systems), knowledge based systems (case-based reasoning and semantic web applications), artificial intelligence and computer vision applications to the analysis of complex and collective phenomena (in particular, crowds of pedestrians). His researches are strongly characterized by an interdisciplinary nature, which implied the collaboration with researchers in the fields of civil and transportation engineering, psychology, social sciences, biology, archaeology and humanities in general, and more recently astrophysics.

• Slides

Ana Bazzan, Federal University of Rio Grande do Sul (Brazil)

• Title: Networks, networks, and more networks: applications in humanities, data science, and machine learning

• Date & Venue: 17/10/2024, 9:30h, URJC Móstoles, Salón de Grados 101, Departamental II

• Abstract:  It is known that networks or graphs can be used in machine learning and data science to represent and analyze data that has complex relationships. Besides, networks are also relevant to the overall AI agenda in at least two aspects. First, it relates to automated data gathering and language models in the semantic web, since the actual data have to be acquired in some manner in order to form the graphs. Second, it can be used to accelerate learning tasks, as in the case of reinforcement learning. In this talk I present examples of how data is acquired and used in applications in the Humanities (history, storytelling) in order to discover patterns and/or to investigate assumptions. Then, I discuss applications on data science and machine learning, as for instance the use of networks in reinforcement learning, with examples from urban mobility and car to infrastructure communication.

• Short Bio: Ana Bazzan is Professor of Computer Science at the Institute of Informatics and Leader of the Artificial Intelligence Group, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil. Her research focuses on multiagent systems, in particular on agent-based modeling and simulation (ABMS), and multiagent learning for the transportation domain. Since 1996, she has collaborated with various researchers in the application of ABMS and game theory to social science domains, such as the emergence of cooperation, the prisoner’s dilemma and public goods games. In recent years, she has contributed to different topics regarding smart cities, focusing on transportation. In 2014, Bazzan was General Co-chair of AAMAS (the premier conference in the area of autonomous agents and multiagent systems).

• Slides

Eugénio Oliveira, University of Porto (Portugal)

• Title: AI & Generative AI: Impact in Society

• Date & Venue: 03/10/2024, 11h, Salón de Grados 101, Departamental II, URJC Móstoles

• Abstract:  We will talk about AI and Generative AI fundamentals, applications impact, challenges, limitations and regulation. The current moment in AI development will be characterized. Gen AI fundamentals and applications will be presented and analyzed together with the main foreseen uses for a Beneficial AI. We also address possible Maleficial AI and the potential dangerous impacts in society. We will emphasize the need of an AI for a sustainable environment versus excessive energy consumption and questionable military and other kind of applications. Emphasis will be made in the need for more explicit legislation on AI development, deployment and use, all around the world. We will finish with few comments on the polemic possibility for an Artificial General Intelligence.

• Short Bio: Eugénio Oliveira is a Full Professor (Emeritus) at the University of Porto, where he created the Distributed Artificial Intelligence & Robotics Group at the Faculty of Engineering in 1989, and was co-founder of LIACC (U Porto's Artificial Intelligence and Computer Science Group). Throughout his career, he has published more than 400 scientific papers mainly in the field of Artificial Intelligence and Multiagent Systems. His main influence has been in directing, collaborating, promoting and supporting activities leading to AI-based applications such as MASDIMA (a Multi-Agent System for Air traffic plan-disruption Management), the ANTE platform for B2B partners selection, FCPortugal RoboSoccer teams, ORBI and ARCA (an Expert System for Cardiac Arrhythmia Diagnosis), that were important breakthroughs at the time of their development. He is now giving seminars on “AI for Social Good”.

• Slides

Ivana Dusparic, Trinity College (Ireland)

• Title: Explainable Reinforcement Learning for Large-Scale Applications 

• Date & Venue: 27/09/2024, 11h, Salón de Grados 101, Departamental II, URJC Móstoles

• Abstract:  RL has seen major breakthroughs in the recent years and is extensively investigated in a range of practical applications, including those within city-scale infrastructures. However, existing algorithms still fall short of being suitable for a wider use in such complex environments. My research focuses on developing techniques that enable the use of RL for optimization in large-scale adaptive systems, for example, communication networks and intelligent mobility. These systems share properties with many other large-scale systems, i.e., are characterized by distributed control, heterogeneity, presence of multiple and often conflicting goals, reliance on diverse sources of information, and the need for continuous adaptation. In this talk I will discuss a range of techniques we have developed for enabling RL use in such environments, such as multi-agent multi-objective optimization, state space adaptation in non-stationary conditions, and online transfer learning. In particular, I will focus on explainability of RL systems (XRL), as a crucial element required for ensuring trustworthiness of RL-based systems. I will discuss differences required in explaining RL systems compared to other types of XAI, and present our approaches to generating and evaluating counterfactual and semi-factual RL explanations. I will conclude the talk by discussing further challenges in enabling RL deployments in large-scale systems, including further development of algorithms to ensure seamless lifelong adaptivity and highlighting the need for scalable explainability, interactive RL, and testing techniques for RL-based applications.

• Short Bio: Ivana Dusparic is an Ussher Associate Professor at the School of Computer Science and Statistics at Trinity College Dublin, Ireland. She has obtained a BSc from La Roche College, Pittsburgh, USA in 2001, and an MSc and PhD from TCD in 2005 and 2010, respectively. Her research expertise is the development of new artificial intelligence algorithms, and specifically reinforcement learning, for optimization of large-scale infrastructures. She has authored numerous peer-reviewed articles in the areas of reinforcement learning agents, multi-agent systems, intelligent mobility, and future communication networks. Her research has been funded through a number of SFI and IRC initiatives, as well as by the industry partners. She is currently TCD lead of the SFI Centre for Research Training in AI, Principal Investigator of the Smart Networking in the Era of AI collaboration between Trinity College Dublin and Tsinghua University, and a Funded Investigator at Enable/CONNECT Research Centre. 

• Slides

Marija Slavkovik, University of Bergen (Norway)

• Title: The Ethics in AI: Perspectives on Fairness, Transparency, and Algorithmic Accountability

• Date & Venue: 16/05/2024, 16h. Salón de Grados 101, Departamental II, URJC Móstoles

• Abstract:  In this talk, we delve into the intricate landscape of AI ethics within the framework of the AI alignment problem, offering a comprehensive exploration of its multifaceted dimensions. Beginning with an elucidation of the imperative to consider ethics throughout the entire lifecycle of AI—from research and development to deployment, procurement, and utilization—we discuss the foundational concepts underpinning AI ethics. Specifically, we unpack the critical components of AI fairness, explainable AI, transparency, and algorithmic accountability, elucidating their significance through real-world examples that underscore the profound implications of technological advancements on society. Drawing upon insights gleaned from everyday scenarios, we underscore the indispensable role of AI experts in comprehending the profound societal impacts of emerging technologies.

• Short Bio: Marija Slavkovik is a Full Professor of Information Science at the University of Bergen where she researches topics on machine ethics (how to automate moral reasoning), ethical issues in user interface design and multi-agent systems. Additionally, in computational social choice, Slavkovik explores Judgment Aggregation and investigates how social network interactions among agents impact collective decision-making. She actively contributes to initiatives aimed at raising awareness and education in AI ethics, including the development of a national PhD course on AI ethics as part of the NORA education committee. She chairs the department of Information Science and Media Studies situated in the Faculty of Social Sciences and is one of Norway's key figures in Artificial Intelligence, representing Norway in various national and European AI research organisations (EuraMAS - European Association for Multi-Agent Systems, EurAI - European Association for Artificial Intelligence).  She is dedicated to strengthening interdisciplinary research cooperation, as well as to education in AI Ethics and presenting her research  in academic as well as in the industry and public outreach venues.

• Slides

Monica Drăgoicea, Technical University of Bucharest (Romania)

• Title: Smart Services, Explainable AI, and smart visualizations

• Date & Venue: 08/05/2024, 10:30h. Room 062, Departamental I. FCEE, URJC Vicálvaro

• Abstract: Today, we have technology that must work for the people’s benefit. We must continue to learn how to intertwine science, technology, and innovation in our fast-growing service-oriented economy, where we deal with people-centric interactions. These interactions are more and more mediated by technology, can be described in complex actor networks, the service ecosystems, and develop in service systems, that integrate people, technology, and shared information. These interactions are more and more "informed" by rich data-driven processes and the citizens are experiencing them more and more through the smart service ecosystems. Therefore, we have the incentives now to better understand, describe, and innovate in complex, service-oriented systems like healthcare, business organizations, government agencies, and cities. And we have the possibility to formulate new value propositions for our society with the new tools that the human mind has created. While AI will always remain artificial because it will always be a product constructed out of the human mind, we cannot deny the potentialities of these new technologies/algorithms to help people do their jobs better or to improve people’s lives. However, there is a growing need for these new technologies to be better explained to increase the degree of acceptance at all levels of society and by all the relevant stakeholders. We want to see a better adoption of AI-type technologies, as this can bring a lot of benefits to society, companies, and individuals. But we need to increase the trust in their decision-building process and create a kind of explainability regarding law compliance, robustness and reliability in operation, and trustworthiness. In this respect, this presentation proposes a journey in the smart service design process, where AI explainability and smart visualization can bring the process of decision-making closer to the most important recipient of this endeavor, the smart citizen.

• Short Bio: Monica Drăgoiceais a Full Professor in Systems Engineering at the Faculty of Automatic Control and Computers at the University Politehnica of Bucharest (UPB). She holds a degree in Electrical Engineering – Automatic Control from UPB (1993), a master’s degree in Engineering Management from Technische Universität Wien, Austria, and the School of Business Administration, Oakland University, Rochester, MI, USA (1999), and a doctorate in Automatic Control from UPB (2000). For the past 25 years, she has been involved in theoretical and experimental work in software and systems engineering. Her research and professional path include several topics, such as artificial intelligence, neural networks, mobile robotics, intelligent control systems for mobile robotics (1993-2006), real-time systems, model-driven development, object-oriented analysis and design (2007-2010), service systems engineering, digital design of services, and computational intelligence, since 2010. Currently, Prof. Drăgoiceais is the Head and founder of the Smart Cities and Robotics iLab, at the Faculty of Automation and Computers from UPB, and the Academic Coordinator of BIS – Business Intelligence Specialist, the Postgraduate Continuing Professional Development Program developed together with SAS Institute.

• Slides

Rosaldo Rossetti, University of Porto (Portugal)

• Title: Artificial Transportation Systems as the basis for studying ITS, Smart Mobility, and Sustainable Communities

• Date & Venue: 07/05/2024, 16h. Salón de Grados 101, Departamental II, URJC Móstoles

• Abstract: In this seminar, we look into the domain of urban mobility from three main dimensions, namely ITS, Smart Mobility, and Sustainable Communities to gain insights into the analytical requirements to support decision making and policy evaluation. We seek to establish an understanding of Artificial Transportation Systems (ATS) from the methodological and architectural perspectives, so as to identify underlying concepts and components, and explore the potential synergies and cross-fertilisation opportunities between ATS and intelligent systems supported by MAS, AI, and ML One aspect of special interest concerns appropriate methods for behaviour modelling and decision-making support.

• Short Bio: Rosaldo Rossetti is an associate professor with the Department of Informatics Engineering and a senior research fellow with the Artificial Intelligence and Computer Science Lab, at the University of Porto, Portugal. He served as a member of the Board of Governors of the IEEE ITS Society during term 2011-2013 and was a member of the steering committee of the IEEE Smart Cities Initiative from 2013 to 2017. He is currently a co-chair of IEEE ITS Society’s Artificial Transportation Systems and Simulation technical activities sub-committee. Dr Rossetti’s primary research interests include behavioural modelling, social simulation, and machine learning with applications to the design of sustainable socio-technical systems. He is also a member of ACM, APPIA (the Portuguese AI Society), and the European Social Simulation Association.

• Slides

Radu-Casian Mihailescu, Heriot-Watt University (UK) & University of Malmö (Sweden)

• Title: Limitations and New Frontiers in Deep Learning and its Applications to Data Science

• Date & Venue: 29/04/2024, 11h. Salón de Grados 101, Departamental II, URJC Móstoles

• Abstract: The advent of deep learning (DL) models has seen neural networks being successfully applied across many classification and prediction problems with implications especially in the areas of image and video, or text and speech processing and analysis. Since its (re-)emergence in 2009 deep neural networks have consistently replaced state-of-the-art approaches to machine learning and largely overcame the problem of hand-crafted feature extraction that dominated to little avail the field precedently, for over 50 years. One of the key enablers of DL models is the availability of compute, yet, although scale has proven to be a definitive driver in building increasingly better performing models, neural nets still exhibit serious vulnerabilities and erroneous behavior in terms of brittleness, spurious correlations, lack of interpretability, or more recently, hallucinating misinformation. In this talk I will set-out by identifying a few key underlying limitations in the deep learning literature, as well as addressing shortcomings of machine learning techniques more broadly. Following, I will point towards several research directions by reflecting on contributions from my previous and ongoing research. The presentation will focus on approaches showcased across various application domains of DL to Data Science problems including computer vision, natural language processing and internet of things settings.

• Short Bio: Dr. Radu-Casian Mihailescu is an Associate Professor in Computer Science at the Faculty Mathematical and Computer Sciences, in Heriot-Watt University (UK). He received a degree in Computing from TU Timisoara (Romania) and a PhD from University Rey Juan Carlos in Madrid (Spain). His research focuses primarily on advances in the field of machine learning (ML), with particular emphasis on state-of-the-art deep learning architectures. Key areas of his research include topics such as out-of-distribution generalization, transfer learning, meta-learning, active learning, interactive learning, few-shot learning, domain adaptation, machine understanding, as well as building upon distributed representations learned by deep networks and incorporating reasoning as an integral part of the learning procedure. Moreover, he is also active in numerous applications of ML approaches to various real-world case-studies such as computer vision, natural language processing for fake news detection, activity recognition based on the internet of things infrastructures, context-adaptive surveillance systems or ambulance coordination for acute stroke care. He is also affiliated with the Internet of Things and People (IoTaP) Research Center at Malmö University in Sweden (MAU), and has previously acted as Program Director for the Applied Data Science Master’s Degree of MAU.

• Slides

Christoph Lütge, Technical University of Munich (Germany)

• Title: AI Ethics -- Opportunities, Challenges & Governance

• Date & Venue: 18/09/2023. Room 170, Departamental II, URJC Móstoles

• Abstract: The increasing presence of artificial intelligence is associated with ethical and governance questions that we are only just beginning to analyze. This presentation explores some ethical opportunities as well as challenges of the use of AI in a variety of fields, including autonomous driving and generative AI. AI ethics fundamentals such as key ethical principles are discussed, as well as initiatives to govern the responsible and safe adoption of AI, in particular the recent EU AI act. 

• Short Bio: Christoph Lütge is Full Professor of Business Ethics at Technical University of Munich (TUM) and the Director of the TUM Institute for Ethics in Artificial Intelligence (IEAI). He is Distinguished Visiting Professor of Tokyo University and has held further visiting positions at Harvard (Berkman Klein Center), Taipei, Kyoto, Stockholm and others. He has a background both in philosophy as well as information studies, having taken his PhD at the Technical University of Braunschweig in 1999 and his habilitation at the University of Munich (LMU) in 2005. In 2007, he was awarded a Heisenberg Fellowship by the German Research Foundation. His most recent books are: “An Introduction to Ethics in Robotics and AI” (Springer, 2021, with coauthors) and “Business Ethics: An Economically Informed Perspective” (Oxford University Press, 2021, with Matthias Uhl). He is a member of the European AI Ethics initiative AI4People and of the German Ethics Commission on Automated and Connected Driving (2016-17). Since 2020, he is Consortium Leader of the Global AI Ethics Consortium. 

• Slides

Other seminars

A full list of invited lectures and previous seminars can be found here.