Correlation and Causality in Responsible AI

This seminar introduces the topic of Responsible AI with a special focus on bias and fairness. It explores the challenge of establishing causal links between key variables and algorithmic accuracy, arguing that current observational techniques fall short for determining causality. An experimental approach based on generative image models is presented as a method for measuring bias, with demonstrations of its application in face recognition systems.

After the seminar, the roundtable discussion, “Disruptive Ideas and Applications in the AI of the Near Future” will follow. It will be moderated by Vittorio Di Tomaso (Managing Director Data & AI – Jakala).

Speakers include:

• Pierluigi Contucci
  Professor of Mathematical Physics, Alma Mater Studiorum at University of Bologna
• Marc Mezard
  Professor of Theoretical Physics at Bocconi University
• Fabio Pammolli
  Full Professor of Economics, Finance, and Data Science – Politecnico Milano, and President, The Italian Institute of Artificial Intelligence (AI4I)
• Pietro Perona
  Allen E. Puckett Professor of Electrical Engineering; Director, Information Science and Technology – California Institute of Technology (Caltech)
  
• Riccardo Zecchina
  Director of the Department of Computing Sciences and Full Professor of Theoretical Physics at Bocconi University

The roundtable offers an opportunity to explore the key challenges and opportunities linked to the evolution of artificial intelligence in the coming years.

Programme

2:30 – 3:00 pm: Registration
3:00 – 4:00 pm: IAS Seminar
4:00 – 4:30 pm: Q&A Session
4:45 – 5:30 pm: Roundtable Discussion

Pietro Perona

Professor Perona’s research focuses on vision: how do we see and how can we build machines that see.
Professor Perona is interested in visual recognition, more specifically visual categorization. In collaboration with his students, he develops algorithms to enable machines to learn to recognize frogs, cars, faces and trees with minimal human supervision, and to enable machines to learn from human experts. His project `Visipedia’ has produced two smart device apps (iNaturalist and Merlin Bird ID) that anyone can download to their smart device and use to recognize the species of plants and animals from a photograph.
In collaboration with Professors Anderson and Dickinson, professor Perona is building vision systems and statistical techniques for measuring actions and activities in fruit flies and mice. This enables geneticists and neuroethologists to investigate the relationship between genes, brains and behavior.
Professor Perona is also interested in studying how humans perform visual tasks, such as searching and recognizing image content. One of his recent projects studies how to harness the visual ability of thousands of people on the web to crowdsource the annotation of images.
Professor Perona is committed to developing responsible artificial intelligence (AI) algorithms. He works on developing experimental methods for assessing algorithmic accuracy and bias in face recognition and other applications of computer vision.