Fysikdoktorandernas seminarieserie hålls via Zoom. En doktorand per tillfälle ger en 15-minuterspresentation om sitt arbete, följt av frågor. Presentationerna riktar sig till fysikintresserade från alla inriktningar.
Zoom link: https://gu-se.zoom.us/j/66921296433
I will present an overview of the research I have done during my PhD studies. I will show the challenges of using standard algorithmic methods for different problems and how I solved them in my research by applying various types of neural network architectures. I found that deep-learning-powered solutions consistently have an increased performance compared to standard algorithmic methods and even simpler machine-learning methods. In particular, I show how deep learning makes novel tasks in image analysis possible and is useful in aiding in the diagnosis of diseases, the prediction of morbidity risks for patients, and in the construction of epidemic strategies.
I am on my last year as a PhD student in the Soft Matter Lab with Giovanni Volpe where I have studied the applications of deep learning in image analysis and medical diagnosis.