DAAD Project Patras School

• 09:30 – 10:00 
  Arrival
• 10:00 – 10:15 
  “Welcome – Introduction”, Thanos Skodras
• 10:15 – 11:00 
  “Responsible Data Science for a Digital Society”, Gjergi Kasneci [1]
• 11:00 – 11:15 
  “Federated Deep Equilibrium Learning”, Dimitris Ampeliotis
• 11:15 – 11:30 
  “AI-Hub @ ECE UPatras”, Kyriakos Sgarbas [2]
• 11:30 – 12:00 
  Coffee Break
• 12:00 – 12:45 
  “How Humans Identify Computer Generated Images and What Their Behaviour Can Tell Us”, Clara Riedmiller
• 12:45 – 13:15 
  “Gaze-Assisted Fake Image DL Discrimination – Experimental Results”, Nikos Fotopoulos
• 13:15 – 14.00 
  Lunch Break
• 14:00 – 14:45 
  “Understanding Student Behaviour in Immersive VR Classrooms Provides Insights for Effective VR Learning Environment Design”, Hong Gao
• 14:45 – 15:15 
  “Proportional Myoelectric Control in a Virtual Reality Environment”, Iliana Loi
• 15:15 – 15:30 
  Coffee Break
• 15:30 – 16:00 
  “Robotized Manipulation of Deformable Materials using Sim2Real techniques”, Nikos Anatoliotakis
• 16:00 – 16:30 
  Discussion and Closing Session

• 10:30 – 11:00 
  Open discussion
• 11:00 – 12:00 
  Visualization and Virtual Reality Group
• 12:30 – 13:30 
  Museum of Science and Technology, University of Patras
• 13:30 – 14:00 
  Open-end discussion and Wrap-Up

[1] Abstract: Data-driven AI algorithms are frequently at the core of digitalization processes. They can be used to identify, optimize, automate and scale solutions to pressing societal problems, such as governance and participation in complex modern societies, personalization of education and health solutions, sustainable planning, production and allocation, and many more. However, for the trustworthy application and long-term adoption of data-driven AI and for a democratic participation in digital processes, challenges around privacy, human-centered transparency, and fairness need to be adequately addressed. This talk focuses on two central questions around human-centered transparency for data-driven AI systems:

• What are the most influential factors for the prediction made by an AI system?
• What can an individual or group of individuals do to achieve a desired outcome?

We will take a look at current approaches to these questions, as well as existing regulatory and technical challenges that can only be addressed in a holistic and principled way by integrating social sciences as a first-class citizen in the devised solutions.