(189bv) An Ordinary Differential Equation Based Machine Learning Framework

Machine Learning (ML) has seen a big resurgence in many areas of engineering, physical sciences, and several areas of research. This presentation will start with a short review of ML strategies such as Feedforward Neural Networks, Deep Learning Convolutional Networks, Self-Organizing Maps and Support Vector Machines. Subsequently the author will present an innovative ordinary differential equation framework for machine learning. The framework allows flexibility with regard to training and validation. The impact of the proposed ML strategy on molecular modeling will be discussed.