Christopher Stover is a doctoral pre-candidate studying pure mathematics at Florida State University in Tallahassee. Focusing on low-dimensional topology and foliation theory, Chris has a general fondness for geometry; even so, his years as a mathematician-in-training have also fostered interest in various other fields, including hypercomplex function theory, functional analysis, and mathematical physics. Chris is entering the third year of his doctoral program and aspires to be a mathematical polymath.
When he’s not attempting to mathematize, Chris does his best to maintain a number of “normal person” hobbies. In particular, he enjoys spending time with his family and friends, traveling, and bearding; he’s also a follower of nearly every sport, an avid music and book collector, and a body modification enthusiast.
Project: Type Inferencing and Predictability
Probabilistic Models and the Wolfram Demonstrations Project
The overall goal of my project is to utilize the type inferencing system designed by Taliesin Beynon and to possibly expand its scope. Ultimately, the object will be to utilize this technology to:
- suggest corrections/improvements to existing code, and to
- implement an “autocomplete”-like mechanism to suggest possible blocks of code completion to Mathematica users in real time.
As I understand it, there are several main components of completing this project:
- Understand and implement Taliesin’s Mathematica notebook importer to scrape all the Wolfram Demonstrations notebooks for inactivated code.
- Once this is done successfully, parse the code and use some sort of machine learning/probabilistic methodologies to assign quantitative values to existing code. The probabilities here represent, among other things, the likelihood that Demonstrations Project authors pass various data types as arguments to the functions therein.
- Utilize these results to complete the tasks outlined in the list above. In particular, interest has been expressed in suggesting corrections/improvements for existing Demonstrations code in the event that Fortran-like code has been utilized; the possibility of implementing this in a sort of real-time autocomplete in future versions of Mathematica was also discussed.
Favorite Outer Totalistic r=1, k=2 2D Cellular Automaton