"I will be investigating how to make deep neural networks more trustworthy, by equipping them with a measure of their own uncertainty. Developing models which can warn us when their predictions are unreliable is crucial if we want to use AI in safety-critical applications like healthcare or autonomous cars. This scholarship is the perfect opportunity for me to delve deeper into machine learning, and I'm excited to be a part of this new academy,” Gala Humblot-Renaux says.
Gala explaines that deep neural networks are algorithms inspired by the human brain which learn to find patterns and make sense of complex data.
“Within computer vision, we specifically work with images – for instance, we can train a neural network to recognize tumors in MRI scans or pedestrians around a self-driving car.”
New academy trains talents
The Danish Data Science Academy is a new national academy that aims at training experts and talents and strengthen interdisciplinary collaboration within data science in a broad sense, including machine learning, artificial intelligence, and Internet of Things. This year they grant 10 PhD scholarships to research projects proposed by visionary and ambitious young researchers, among them Gala Humblot-Renaux.
Each PhD is associated with a supervisor at a Danish university. At Create, Professor Thomas B. Moeslund will be Gala’s supervisor. He is a leading researcher within AI in Denmark and a part of the national Pioneer Centre for AI.
“Both in our research group, Visual Analysis and Perception, and at the Department for Architecture and Media Technology, we are very honored that Gala has received a PhD grant from DDSA. The research problem Gala will be working on is situated at the very core of modern AI. While current AI algorithms are powerful in their predictive power their uncertainties are still poorly understood, which hinders widespread use of such algorithms in society. The work Gala embarks on can have a profound impact on society by paving the way for smarter and better technologies within all fields ranging from personalized medicine to intelligent robots,” states Thomas B. Moeslund.