James Pedersen is a rising junior studying pure mathematics (and possibly computer science in the future) at the University of Washington in Seattle. In the fall of 2017 he did research concerning an application of Brownian motion to options trading. Although he intends to pursue graduate study in mathematics, he is also interested in robotics and machine learning (in particular, defending neural networks from adversarial examples). He loves to program and is willing to work on practically any programming project that pushes him to learn new things. In his spare time he enjoys playing the piano (particularly Chopin) and running.
Project: On the Road to a Gravitational Computer: Encoding Basic Programs in Terms of the Three-Body Program
To find methods by which basic programs, such as addition and multiplication, can be encoded in terms of the (Newtonian) three-body problem.
Main Results in Detail
I began work on an algorithm to transform a known spiral-like solution to the three-body problem into a form (like that in the image) that encodes the multiplication of two numbers.
- 1. Find stable spiral-like solutions to the three-body problem where the three bodies eventually undergo a triple collision.
- 2. Train a neural network to encode any computer program in terms of the three-body problem.