Welcome to The ADAM Project’s documentation!¶
ADAM is a python-based project that leverages the stochastic gradient descent algorithm ADAM to optimize nuclear systems. This project is a part of Noah Walton’s Ph.D. work and is collaboratively developed by members of Dr. Vladimir Sobes’ research group at the University of Tennessee. The ADAM stochastic gradient descent algorithm in it’s most general form can be found here. The methodology as it applies to nuclear systems is inspired by the ability to calculate stochastic sensititvies when solving Monte Carlo neutron transport calculations. These sensitivities can be converted to derivatives and used as input to a gradient descent algorithm, or in this case ADAM. The purpose of this python package is to couple a Monte Carlo neutron transport solver to the execution of the ADAM algorithm and automate the step-by-step gradient descent. A scientific journal article that details three challenge problems solved by coupling ADAM to Monte Carlo sensitivities can be found here.
The following sections will describe how to install and use ADAM. Also, the API give more documentation on the internal source code for development and expansion of capabilities.