Projects
-
A self-directed project to further my understanding of reinforcement learning using Python. Neural networks play games against each other and learn how to play the games. Tic-tac-toe is the starting game, but I’ll add checkersClick here as well.
The neural networks will be based on tensorflow to ensure that I don’t cross over into material covered by school courses and to show my own mastery of the library.
Click here to see the work in progress.
-
Using PySpark to create a neural network analytical tool, we correlated the Social Vulnerability Index with an area’s public/private school ratios, graduation and dropout rates, and teen employment rates.
Click here to see the repository.
-
The idea was to see how I could tune the performance of a classification model by making a quick decision that would eliminate most of the classification options and then usClick heree a more specialized model to delineate between the remaining classification options. I ended up using the MNIST data set as a proof-of-concept and would like to continue working on this in my free time to see how well it does on more complicated data sets.
Click here to see the Jupyter Notebook.
-
I researched how the long short-term memory mechanic could map to the function of brain cells.
Click here to go to the draft of the paper as of December 5, 2021.
-
The final assignment from my introduction to Machine Learning class. I used the Reuters Corpus Volume 1 dataset from scikit-learn and compared Decision Tree classifiers with a Neural Net classifier. I did not then have the experience to attempt the large project in a distributed system like Spark or Hadoop at the time, but am working on learning those things.
Click here to go to the Jupyter Notebook
-
A group project in which we took two Raspberry Pi devices and connected them over the internet. One device transmitted telemetry data (which way the device was “looking”) and the other device read the telemetry data and displayed it on a physical sphere.
Click here to read the project report.
-