Vision is central to how humans navigate the world, whether recognizing a familiar face in a photo or driving to a family dinner. For Artificial Intelligence (AI), however, even minor visual distortions, such as changes in brightness, contrast, or subtle perturbations, can cause object recognition algorithms to fail. Bridging this performance gap has been a major challenge in machine learning.

Host

Klaas van Gisbergen, PhD, Tissue Immunity Lab

Venue

Seminar room

As we were wrapping up an earlier conversation with Champalimaud Foundation (CF) Clinical Director Professor António Parreira, I mentioned that there would be a sister article featuring one of his colleagues, Joe Paton, Director of Neuroscience Research at CF. The plan was to ask the same questions, more or less, to explore whether the cultural outlooks of the clinical and research branches aligned after 20 years. I invited António Parreira to open Joe’s interview with a question.

When António Parreira joined the Champalimaud Foundation (CF) more than a decade ago, the place was still more vision than institution. The building stood ready, white limestone gleaming by the river, but inside, there was little activity. “When I arrived, there were just two of us – two doctors,” he recalls. “It was too small then to think of a community.”

The Champalimaud Foundation (CF) has the advantage of housing a research institute and a clinic under the same roof. Collaboration opportunities are everywhere, and from the start Henrique Veiga-Fernandes made the most of them. He's been linking research and clinic, collaborating with health professionals, and hopes to expand these while inspiring others to do the same.

When Maria João Cardoso arrived at the Champalimaud Foundation (CF) in 2011, the building was almost empty. “There was almost no one there”, she recalls. “It was a ghost institution that progressively came to life”.