DISTINGUISHED SPEAKERS

Bridge

Dr Björn W. Schuller

Professor
Imperial College
London, UK
Bridge

Prof. Dr. Ling Liu

Georgia Institute of Technology
IEEE Fellow
Editor in Chief, ACM Transactions on Internet Technology

Bridge

Dr. Keeley A Crockett

Professor
Manchester Metropolitan University
UK


Bridge

Dr. Zhongfei (Mark) Zhang

Professor
Binghamton University
USA
Bridge

Dr. Roger B. Dannenberg

Emeritus Professor of
Computer Science
Art & Music
Carnegie Mellon University
USA

Bridge

Dr. Seán McLoone

Director
Centre for Intelligent and Autonomous Manufacturing Systems
Queen's University Belfast
UK
Bridge

Dr. Jean-Paul Van Belle

Professor
University of Cape Town
South Africa
Bridge

Prof. Dr. Jan Treur

Professor
Vrije Universiteit Amsterdam
Netherlands
Bridge

Dr. Christian Blum

Senior Research Scientist
Artificial Intelligence Research Institute (IIIA)
Spanish National Research Council (CSIC)
Spain
Bridge

Prof. (Dr.) Alfredo Cuzzocrea

Director of The Big Data Engineering and Analytics Laboratory (iDEA Lab)
University of Calabria
Italy

Bridge

Dr. Hajer Baazaoui Zghal

Professor
CY Cergy Paris University
France



Bridge

Prof. Sandhya Samarasinghe

Professor
AI and Complex Systems Modelling
Lincoln University
New Zealand
Bridge

Dr. Hari Mohan Pandey

Professor
Data Science and Artificial Intelligence
Bournemouth University, UK
Bridge

Dr. Ashish Seth

Professor
Inha University Tashkent
Uzbekistan



Bridge

Dr. Mario Jose Divan

Sr. Solution Architect and Service Lead
Intel corporation
Bridge

Dr. Ruggero Donida Labati

Professor
University of Milan
Italy
Bridge

Dr. Maia Angelova

Professor
Al in Health at Aston Digital Futures Institute
Aston University
Birmingham, United Kingdom
Bridge

Dr. Gautam Srivastava

Brandon University
Canada




Bridge

Dr. Angelo Genovese

University of Milan
Italy
Bridge

Dr. Marita Turpin


University of Pretoria
South Africa
Bridge

Prof.(Dr.) Keith Sherringham


PMO Specailist
New South Wales
 Australia


Bridge

Dr. Wisam Bukaita


Professor
Lawrence Technical University
USA
Bridge

Dr.(Smt.) N. Ranjana

Director
Directorate of Systems & Technology
Defence Researh & Decvelopment organization(DSTA)

Bridge

Dr. Prabhat Manocha

Associate Director
IBM
Bridge

Dr. Raghu Ramakrishnan

Chief Architect & Technology Head
Government Business Unit
Tata Consultancy Services


Title of the Talk: Health Informatics: The Data Awakens

×
Abstract : Massive data is currently being collected from various personal devices worn and carried with us 24/7. Such data includes the interaction with our phone, but also audio, video, and a rich plethora of biosignals coming from smart watches and smart jewelry such as rings. While this data is already often targeted for health applications, there is dormant massive potential on a much larger scale to exploit it. Here, we shall discuss doing so ideally already for prevention of illnesses. To this end, deep learning approaches are presented able to deal with various types of data by end-to-end learning and neural architecture search. In addition, tapping into large models and potentially emergent abilities of these in the health domain will be featured. Overall, tomorrow’s smart devices might improve health and save lifes of millions once the data awakens the right way.

Title of the Talk: People Powered AI – Challenges and opportunities in Responsible AI Development.

×
Abstract : The operalisation of ethical principles, current and emerging legalisation, and the understanding and mitigation of potential consequences to individuals and society are key challenges in the design, development and deployment of Artificial Intelligence (AI) driven systems. Organisations developing AI solutions as a service or innovating novel applications will need to openly address ethical principles such as bias, fairness, explainability, transparency, data privacy, accountability, and safety through AI Governance. In this talk I will briefly first overview ethical issues and emerging/current legalisation related to AI systems and the impact on people and society. Responsible innovation and ethical tech are essential to build public trust and meet legal obligations, yet how can businesses and researchers in academia harness People Power through participatory AI (co-design, co-production and public engagement). I will give some examples of research projects that bridge the gap between businesses, academia, and people and how the use of ethical toolkits such as consequence scanning and harms modelling that can help build public confidence and how enhance how academics undertake research.

Title of the Talk: Addressing Reliability for Deep Learning

×
Abstract : One significant obstacle to deploying deep neural network (DNN) models in real-world applications is that deep learning systems often break down in novel situations. Specifically, DNNs tend to yield unreliable predictive uncertainty estimates and make high-confident yet incorrect predictions when exposed to inputs drawn from unfamiliar distributions. Consequently, accurate predictive uncertainty analysis of DNNs is critical in many high-stake applications such as medical diagnosis, self-driving vehicles, and financial decision-making, where silent mistakes can lead to catastrophic consequences. In this talk, I will first introduce a novel uncertainty factorization model as a theoretical foundation for uncertainty analysis in general. Based on this model, I will then introduce a general and flexible framework for predictive uncertainty estimation with promising evaluation results in several out-of-distribution detection tasks on both vision and language datasets.

Title of the Talk: AI and Music: Why Can't Computers Learn to Write Songs?

×
Abstract : After years of effort, data-driven deep learning has failed to learn how to create convincing music. Although there are very impressive results, deep learning systems seem far from creating even basic songs that are well-organized and whole. In contrast, rules and knowledge engineering have been used to create very convincing songs, but are unable to learn or invent music that is highly original. I will examine human-composed music and AI music generation through the lens of entropy and repetition, shedding light on how AI music systems diverge from human composers. Our findings offer challenges for future music learning systems and offer new ways to evaluate music generation efforts.

Title of the Talk: Generative AI and the potential of open-source Large Language Models

×

Title of the Talk: Modeling Multiple Orders of Adaptivity Using a Higher-Order Adaptive Dynamical System Modeling Approach

×
Abstract : Realistic processes are usually considered to form complex dynamical systems. Often such dynamical systems are adaptive; they can be conceptualised and modeled by adding an adaptation process on top of the base process. As a common practice, this is modeled as a kind of adaptation algorithm, like a learning algorithm or progam code for it. However, in many cases the adaptation process itself can also change over time in a context-sensitive manner. This introduces a distinction between first-order adaptation (adaptation of the base process) and second-order adaptation (adaptation of the adaptation process). Moreover, as in some cases also a second-order adaptation process can themselves be adaptive over time, depending on changing contexts, even more orders of adaptation are considered. In this keynote speech, different examples of such orders of adaptation will be discussed, with inspiration taken from different scientific domains. These examples vary from plasticity and metaplasticity considered in neuroscience (first- and second-order adaptation) to still higher orders of adaptation considered in evolutionary biology (up to third- or fourth-order adaptation) and higher orders of adaptation considered in genetics and epigenetics and their effect on mental disorders (up to fifth-order adaptation). When adaptation is modeled like an algorithm or even a piece of program code as is common practice, it is not easy to build higher-order adaptation on top of it in a neat and transparent manner: one would have to consider algorithms adapting other algorithms, or pieces of program code adapting other pieces of program code. To address this difficulty, it is shown how a network-oriented modeling approach based on self-modeling networks can be used. In that case, a base dynamical system is modeled as a base temporal-causal network model and any next order of adaptation of it is modeled in the format of such a network too by applying the self-modeling principle for networks, also called the network reification principle. This provides a neat and transparent declarative description for all desired orders of adaptation within one overall temporal-causal network model according to different levels of self-modeling. It is shown that any smooth (higher-order) adaptive dynamical system can be modeled like this. This approach will be discussed in some detail and it will be illustrated for the different examples of higher-order adaptivity up to fifth-order adaptation as mentioned.

Title of the Talk: STNWeb: A Web Application for Visualizing Optimization Algorithm Behavior

×
Abstract : STNWeb is a recent web-based tool for the visualization of the behavior of optimization algorithms such as metaheuristics. It allows for the graphical analysis of multiple runs of multiple algorithms on the same problem instance and, in this way, it facilitates the understanding of algorithm behavior. It may help, for example, in identifying the strenghts as well as the weaknesses of an algorithm. In this way it can help the algorithm designer to iteratively adapt and change the algorithm in order to improve its performance. STNWeb is designed to be user-friendly. Moreover, it is offered for free to the research community. In this talk I will give an introduction to STNWeb, and I will talk about latest improvements and extensions, including the use of large language models (LLMs) for an automized interpretation of the generated graphics.

Title of the Talk: Temporal Big Data Analytics

×

Title of the Talk: Machine Learning-based Ontology Enrichment

×

Title of the Talk: Artificial Intelligence and Health

×
Abstract : Human health is a complex domain where life changing decisions are made daily by health specialists all over the world. These are challenging decisions made under uncertain conditions of incomplete Knowledge and understanding of human health in general, and response of an individual to a health challenge in particular. Artificial Intelligence (AI) can be a valuable tool and aid in these decision makings. In this talk, I will go over some of our current and past research on AI applications in health covering a spectrum health challenges and solutions. These encompass a range of AI methodologies including neural networks and machine learning to diagnose diseases, find therapeutics and understand and model with AI biological mechanisms underlying health and disease in the case of number of health challenges including breast cancer, Alzheimer’s disease, Tuberculosis and biological regeneration.

Title of the Talk: Drone-camera based human activity recognition (HAR) using Deep Learning Model

×

Title of the Talk: Smart Infrastructure for effective multi-cloud collaboration

×
Abstract : Currently, many organisations are working across multiple collaboration platforms. For some, it’s to support business continuity, or because they’re in the process of gradually migrating from a legacy platform. For others, it’s about meeting the needs of specific user groups or connecting their front- and back-office workers.In many cases, collaboration tools help them build a hybrid working strategy that can flex to meet the changing needs of their organisation. However, multiple tools can create complexity that’s difficult to manage and maintain, requiring on-demand user training, support and configurations. Switching between platforms can make communication and collaboration more time consuming, and calling charges can make this approach more expensive. Living and working in a multi-cloud world shouldn’t have to disrupt voice services or collaboration. Ideally, what users want and need should be the priority, and the tech should follow their lead, providing organisations with flexible infrastructure and solutions that adapt. This talk will explore various possible infrastructure and areas to help build a successful multi-cloud collaboration strategy.

Title of the Talk: Reliability Monitoring on Edge Computing

×
Abstract : System reliability monitoring focuses on determining the level at which the system works as expected (under certain conditions and over time) based on requirements. The edge computing environment is heterogeneous and distributed. It may lack central control due to the scope, number, and volume of stakeholders. Multiple business use cases throughout different verticals run distributively, such as object detection-based or health applications. The artificial intelligence models are valuable and advantage-competitive in these solutions, producing different performances and optimizations. This talk discusses the context and scenarios related to edge computing jointly with their particularities. After that, it identifies and characterizes the main efforts related to System Reliability Monitoring strategies using Artificial Intelligence models for supporting decision-making processes. Next, challenges, opportunities, and trends are discussed with main remarks. Finally, it discusses some conclusions about the ecosystem and trends.

Title of the Talk: Artificial Intelligence for Biometric Technologies

×
Abstract : The number of biometric applications and devices is continuously growing on global scale and biometrics is pervasively entering the everyday life of users. This relevant expansion is producing new challenges and requirements to be fulfilled by the designers. Features such as adaptability, enhanced interactions with the user, robustness to non-ideal conditions, real-time capability, privacy-compliance, and high accuracy are strongly required in innovative applications and solutions, such as: cyber security, smart devices, and ambient intelligent infrastructures. The talk will focus on innovative biometric recognition approaches and systems, with specific attention on recent approaches based on artificial intelligence and deep learning techniques. Artificial intelligence methods and deep learning approaches are capable to learn discriminative features directly from complex multidimensional signals and increase the accuracy, adaptability, and robustness to non-ideal conditions of biometric systems with respect to traditional approaches. The talk will present biometric systems from a technological point of view and will provide an excursus of recent artificial intelligence approaches, including deep learning methods with current strong points and limitations.

Title of the Talk: AI for precision healthcare

×
Abstract : Precision healthcare is an emerging approach for personal healthcare and disease prevention that considers the individual’s variability in health, genes, lifestyle and environment. In this talk, I will show our latest research and describe how our data-driven models, based on physical and physiological data, can assist in the early detection of insomnia, diabetes and chronic low-back pain, and provide a robust basis for pre-screening of individuals with wearable devices. Furthermore, I will show how the explainability of such models can be enhanced by combining them with domain driven knowledge and hybridization. I will consider the following case studies based on our most recent projects: (a) Glucose-insulin regulation models: glucose-insulin dynamics is central for understanding the regulation mechanisms between different organs in the human body and is key to maintain healthy life and prevent diabetes. We combine dynamical systems approach with machine learning algorithms to model the regulation between glucose and insulin and predict glucose dynamics and insulin utilisation in healthy and pre-diabetic regimes [1,2]; (b)Classification of chronic back pain from multi-modal data: we use image, questionnaires, and physical activities data to classify chronic low back pain [3,4]; (c) Classification of insomnia and sleep from physical activities(actigraphy) data. Our models can distinguish acute from chronic insomnia and healthy sleep [5-8]. They provide a robust basis for pre-screening of insomnia with wearable devices in a home environment.

Title of the Talk: Federated Learning for Data Privacy

×
Abstract : In recent years, mobile devices can be equipped with increasingly advanced computing capabilities, which opens up countless possibilities for meaningful applications. Traditionally, any cloud-based Machine Learning (ML) approach requires that data be centralized on a cloud-based server/data center. However, this can result in critical issues related to unacceptable latency and communication inefficiency as well as major security and privacy concerns. However, conventional ML technologies still require personal data to be shared. Recently, in light of increasing security and privacy concerns, the concept of Federated Learning (FL) has been introduced. In FL, end devices use their local data to train a local ML model required by the server. In a large, complex mobile edge networks, FL still faces implementation challenges with regard to communicational costs, resource allocation, security, and privacy. In this talk, we begin with an introduction to the background and fundamentals of FL. We then discuss how FL can work to try and preserve privacy while maintaining security of data. Finally, we discuss some open research areas and specific open problems where attendees may be able to make an impact.

Title of the Talk: Deep Learning for Hematopathology

×
Abstract : Computer Aided Diagnosis (CAD) systems are increasingly utilizing image analysis and Deep Learning (DL) techniques, due to their high accuracy in several medical imaging fields, including the detection of Acute Lymphoblastic (or Lymphocytic) Leukemia (ALL) from peripheral blood samples. CAD systems based on DL can support the pathologists in performing their decision by analyzing the blood samples images to determine the presence of lymphoblasts. However, when using DL, the limited dimensionality of ALL databases may highlight bias in the data and cause overfitting, favoring the use of transfer learning techniques to reduce the bias and increase the accuracy in the detection, in particular by considering Convolutional Neural Networks (CNN) and Vision Transformers (ViT) pretrained on larger databases. This talk will present recent possible solutions for high accuracy ALL detectionbased on CNNs and ViTs, addressing some of the problems of medical data, such as bias and limited dimensionality.

Title of the Talk: Prompt Engineering when using Generative AI

×

Title of the Talk: Cloud-Based Project Services with Artificial Intelligence in Business Transformation

×
Abstract : The pace, extent, complexity, impacts, and risk from business transformation are increasing. With business transformation often delivered via projects (alongside the need to sustain operations), there is an increasing premium on Project, Program, and Portfolio management. A corresponding need for Project Management Offices (PMOs) and emerging Transformation Management Offices (TMOs) operations is also occurring, together with a demand for cloud-based project services like Project Management as a Service (PMaaS) or Project Management Office as a Service (PMOaaS). Beyond systems integration and data assurance, cloud-based project services have a dependency upon stakeholder collaboration, related service management, and business adoption and adaption. Artificial Intelligence (AI) is being increasingly folded into cloud-based project services (like e-discovery, dependency management, document management, contract reviews, resourcing, text library services, analytics, and reporting). AI integration into controls (budget, scope, resources, schedule), is also occurring. Emerging aspects of AI and cloud-based project services are considered in this session, with some related business opportunities.

Title of the Talk: Global Warming's Influence on Temperature Increase

×
Abstract : This study utilizes advanced computational tools and statistical methods to analyze three-term patterns in temperatures from 1900 to 1991, from 1991 to present, and forecasts for the next ten years. A combination of algorithms, coding, and machine learning techniques are employed to visually represent the behavior of temperature trends, incorporating a 95% confidence interval and smoothing coefficients Alpha, Beta, and Gamma. The data reveals a consistent increase in temperatures over the past 30 years, with a noticeable variation in temperature distribution throughout the calendar year compared to the previous century. Of particular interest are the projected rises in average temperatures for the months of April, June, August, and October, with July expected to experience the highest average temperature. Additionally, a striking observation is the significant difference in average temperatures for the month of January, potentially indicating the exclusion of this month in the temperature's calendar. Looking ahead, the forecasted data shows a pattern of steadily rising average temperatures, surpassing those recorded in the past three decades. This progression also suggests a potential merging of the average temperatures for January and December towards that of February, suggesting further warming trends and potential ecological impact.