This is a series of notes that I prepared during my job seeking in the last few months and there will be four separate notes in total:
This note will cover some very basic concepts of machine learning and will focus on the application, instead of the theory.
Mathematica (MMA) is known for its comprehensive library that covers almost everything you need in mathematics, physics and engineering. While I’m a big fan of MMA and MMA released its machine learning (ML) features long time ago, I used to use Python for ML practices because Python is more widely accepted by the data science community and thus, provides better supports. However, it has been 4 years after MMA 11’s first release, the ML features in MMA are getting much better than before. So, I’ve spent sometime playing with MMA ML and made such a reference for later conveniences.
I came cross the work Creation and manipulation of quantized vortices in Bose-Einstein condensates using reinforcement learning a few months ago and found it quite interesting.
I’ve been thinking about a class of optimization problems on a quantum computer and wish to explore them with machine learning algorithms, especially reinforcement learning, as it does not require a good initial guess (radom, intuition or whatever). More recently, I came across to the paper “Global optimization of quantum dynamics with AlphaZero deep exploration”, which has been published online about 2 weeks ago. While the authors considered a rather ideal situation, their work did shed lights on my research. So, I prepare this note to briefly discuss their work and some relevant topics.
