LEARNING PHYSICS WITH NEURAL NETWORKS

Manager:  William Zhu – In-house

Contactwilliamzhu@g.ucla.edu

Learning Objectives:  Programming (Python, Keras, and Tensorflow), Neural Networks, Deep Learning

APPLICATIONS: OPEN

The goal of this project is to train a deep learning model to predict future states of a closed system based on a known sequence of past states, in a physics-consistent fashion. We will build and run simple physics simulations, then assemble obtained data into an appropriate dataset, on which we will train our network.