Program 2023

After the conference, the authors and participants are cordially invited to the social event "Artificial Intelligence at the Pub" at 19:00. This is a pre-event of the partner conference AI+.   

Keynotes:

David Doermann, Professor at the Department of Computer Science and Engineering, University at Buffalo (USA)

Title: The Hidden Dangers of Generative Models: Uncovering the Risks and Mitigating the Harm

Abstract: Generative models, such as GANs, VAEs, and language models, have revolutionized the field of AI and transformed our ability to generate high-quality images, videos, and text. However, these powerful tools also come with hidden dangers that can have serious consequences for individuals and society as a whole. In this keynote talk, we will explore the risks associated with generative models and the potential harms they can cause, including privacy violations, algorithmic bias, and disinformation. We will examine case studies where generative models have been used for malicious purposes, and discuss the ethical and social implications of these actions. Additionally, we will examine current research efforts to mitigate the risks of generative models, including strategies for detecting and preventing harmful content, ensuring transparency and accountability, and promoting responsible use. By understanding the hidden dangers of generative models and developing effective mitigation strategies, we can ensure that these powerful tools are used for the greater good and not for harm.  (This abstract was written entirely by ChatGPT)

Bio: Dr. David Doermann is a Professor of Empire Innovation and the Director of the Insitute for Artificial Intelligence and Data Science at the University at Buffalo (UB). Before coming to UB, he was a Program Manager with the Information Innovation Office at the Defense Advanced Research Projects Agency (DARPA). At DARPA, he developed, selected, and oversaw research and transition funding in computer vision, human language technologies, voice analytics, and media forensics. From 1993 to 2018, David was a research faculty member at the University of Maryland, College Park. In his role in the Institute for Advanced Computer Studies, he served as Director of the Laboratory for Language and Media Processing and as an adjunct member of the graduate faculty for the Department of Computer Science and the Department of Electrical Engineering. He and his group of researchers focused on many innovative topics related to the analysis and processing of document images and video, including triage, visual indexing and retrieval, enhancement, and recognition of visual media's textual and structural components. His recent research has focused on advanced AI techniques applied to computer vision, medical image analysis, federated learning, neural architectural search, binary neural networks, and the detection of false and misinformation in multimedia content. David has over 300 publications in conferences and journals, is a fellow of the IEEE and IAPR, has numerous awards, including an honorary doctorate from the University of Oulu, Finland, and is a founding Editor-in-Chief of the International Journal on Document Analysis and Recognition. 

Massimiliano Ruocco, Senior Research Scientist at SINTEF and Adjunct Associate Professor at NTNU, Trondheim (Norway) 

Title: Transforming Time Series Analysis: Exploring the Potential of Transformer Architecture for timestamped data

Abstract: Time series analysis is an important area of research in various fields such as finance, economics, meteorology, and signal processing. With the increasing volume of time series data generated by various systems, the demand for effective and accurate time series analysis methods has become more crucial. Traditional time series analysis techniques have proven to be effective in modeling time series data, but they are often limited in their ability to capture complex patterns and dependencies that may exist within the data.
The transformer architecture, initially introduced for natural language processing tasks, has shown remarkable success in various other domains, including image processing, speech recognition, and audio synthesis. Recently, researchers have explored the possibility of using the transformer architecture for time series analysis over different machine learning tasks.
This presentation provides an overview of the transformer architecture and its potential for time series analysis. We discuss the pros and cons of using transformer-based models for time series analysis, including their ability to capture long-term dependencies, handle irregularly sampled data, and deal with missing data. Finally, we present some real-world examples where transformer-based models have been applied in various time series analysis tasks, including forecasting, anomaly detection, and classification. Overall, this presentation aims to shed light on the potential benefits and drawbacks of using transformer architecture for time series analysis and provides a roadmap for future research in this area.

Bio: Dr. Massimiliano Ruocco is Senior Research Scientist at SINTEF Digital at the Department of Software Engineering, Safety and Security in Trondheim (Norway) and Adjunct Associate Professor at the Norwegian University of Science and Technology in Trondheim (Norway). Massimiliano is associated with the Data and Artificial Intelligence Group at the NTNU and a research member of the Norwegian Open AI Lab. Massimiliano is project manager and principal investigator in several projects financed by the Research Council of Norway, including Machine Learning for Irregular Time Series (ML4ITS) and LandSkape - Hybrid Physical-Based Deep Learning for Fast and Reliable Wind Flow Estimation. Massimiliano has a strong background in AI and ML, having worked in both industry and academia. He has experience in both the theoretical aspects of ML and its applications. Currently, Massimiliano is focusing on modern AI techniques for time series analysis, particularly on supervised, self-supervised and foundation models for the analysis of temporal sequences.