Professor Jaakko Lehtinen of Aalto University and FCAI talked with the Finnish public broadcaster Yle on the adversarial neural networks developed with his research team at Nvidia. The networks Lehtinen and his colleagues created were able to create – on their own – life-like but artificial human faces by competing against each other at recognising whether a certain image is real or fake. The networks used a sample set of thousands of actual images of celebrity faces and then gradually went on to create convincing artificial faces.
The story online (in Finnish): https://yle.fi/uutiset/3-10115902
Lehtinen appeared also in the main news broadcast at Yle TV 1 yesterday talking about the creativity of algorithms. Also interviewed was Aalto University's Head of Research Strategic Support Ella Bingham, a PhD and Docent in data analysis and artificial intelligence herself. Bingham talked about what the top AI researchers in Finland are doing within FCAI, for instance, improving the mutual understanding of humans and machines.
News insert, starting at 10.20 (in Finnish): https://areena.yle.fi/1-4234984