MICCAI 2020

3 Alaa Bessadok 9 DAILY Mo n d a y Translating this work from the paper to the real world is about building strong collaborations with hospitals and doctors, and national and international bridges , Islem tells us. Such powerful intelligence requires a great deal of data , but once you have it, you can stop collecting and start predicting. “ I always say believers are achievers ,” Islem grins. “People who believe in this intelligence. People from different disciplines joining forces and building synergies which will translate into the real world. Clinicians, doctors, radiologists, neurologists, people who manage hospitals. You name it. I do not see this as a one-person job. We are not very far away from this .” Islem’s team have already pioneered many models for designing these predictive intelligence synthesizing time-dependent brain graphs. Islem has counted for her group no less than 12 papers at MICCAI this year , including 9 workshops, most of them revolving around designing this predictive intelligence. She hopes that many other research teams around the world will want to build on their work and thinks they will make an impact within the next five years once they have good predictive accuracy . Even without very high accuracy, doctors can consider the confidence of the prediction and the accuracy level of the model itself while they are doing their diagnosis. They do not need to wait for 100 per cent accuracy or 99.9 per cent accuracy. It could start at 75 per cent, if the doctors and radiologists are aware of it. “It is like asking a doctor for their opinion,” Islem explains . “You are asking what they think about the diagnosis of a patient. You know that the doctor is good but not outstanding. This is exactly what AI is doing. You are asking the AI system to show you how to design a more personalized treatment for a patient. The AI system will respond that it I can help you, but its confidence is 70 per cent and its accuracy is 75 per cent. We need to start somewhere .” Finally, we ask Alaa, what has she learnt from working with Islem? “I learned how to start a research project and really think about a problem, and how to read, write and present research papers . I learned how to criticize existing works so that I can come up with new ideas and solve existing limitations . I have learnt a lot from Islem – and I am still learning from her!” To find out more about Islem and Alaa’s work, visit their oral which is part of Neuroimaging A today (Monday) at 11:30 (UTC). Awesome Islem was also featured as Woman in Science on Computer Vision News of October 2020.