The combination of machine and deep learning and medical records have the potential to improve decision making in medicine, and may lead to significant improvements in patients’ quality of life and the economics of medical organizations. Avi shoshan will present the main hurdles and problems in working with large datasets of patient records from different sources, a pipeline of tools built to handle those problems, and actual modeled examples built with it. Problems such as efficient ways to hold the data, cleaning, normalizing, complex feature generators, modeling, data bias, and performance testing will be discussed combined with real life examples from databases of millions of patient records.
Avi Shoshan has +20 years’ experience in Algorithms, Software Development, Data Exploration, Genomics and Image Processing. At Medial EarlySign, Avi is charged with developing algorithmic infrastructure, machine and deep learning models, and the technological architecture of several R&D projects. Prior to EarlySign, Avi was VP R&D at Emza, where he developed real time video analytics algorithms running on embedded devices. He also developed algorithms for high end storage systems at Kashya & EMC. Furthermore, Avi headed a research group focusing on bioinformatics and computational genomics research at Compugen. Avi holds a BSc in Math & Computer Science from the Hebrew University, and an MSc in Computer Science from Tel Aviv University. He published several papers and submitted patents in the fields of algorithms and genomics.