Public defence in Computer Science, M.Sc. Yogesh Kumar

Title of the thesis: Efficient Transfer Learning with Sequential and Multi-Modal Approaches for Electronic Health Records

Doctoral Student: Yogesh Kumar
Opponent: Professor Line Clemmensen, University of Copenhagen, Denmark
Custos: Associate Professor Pekka Marttinen, Aalto University School of Science, Department of Computer Science

This thesis explores how deep learning can be used to improve the analysis of Electronic Health Records (EHR). EHRs contain valuable patient information that can be used to predict healthcare needs and outcomes, but the complexity and quality of the data often pose challenges. The research aimed to develop Artificial Intelligence (AI) models capable of making more accurate predictions while addressing issues such as limited data availability and the need for expert input to ensure data quality. 

The study developed new techniques to enhance the predictive power of AI models, improving accuracy of healthcare demand forecasting. By utilizing transfer learning, the models were able to perform effectively even when working with smaller datasets, a common limitation in healthcare data. Additionally, the research developed new methods for improving how different models can be compared and for the integration of expert knowledge into medical imaging models, enhancing their accuracy and efficiency.

The results of this study offer practical benefits for healthcare providers by helping them make more accurate predictions, optimize resources, and deliver better care to patients. These advancements set the stage for future developments in healthcare analytics, bridging the gap between cutting-edge AI research and real-world medical applications.

Thesis available for public display 10 days prior to the defence at: https://aaltodoc.aalto.fi/doc_public/eonly/riiputus/ 

Doctoral theses of the School of Science: https://aaltodoc.aalto.fi/handle/123456789/52