Machine learning has been one of the main success stories of computer science over the last few decades. Today, the cutting edge of machine learning is deep learning, and deep learning has been key to designing intelligent systems that can leverage big data to address a host of engineering applications ranging from computer vision, to robotics, to natural language understanding, and to speech recognition.
We will present our development and application of deep learning methods for problems in the natural sciences including: (1) the detection of exotic particles in collider data; (2) the prediction of the physical, chemical, and biological properties of small molecules; (3) the prediction of chemical reactions; and (4) the prediction of the structural features and 3D structures of proteins.