Skrusty Learn
- Ishaan Gupta (ishaang5)
- Kyubin Min (kyubinm2)
- Jordan Smith (js149)
- Aaren Kang (aarenk2)
We are attempting to create an AI library in rust akin to scikit-learn, helping users train AI models in rust. We will implement everything from scratch, starting from tensor processing building up through AI training algorithms and concluding with fully-fledged configurable AI models. We decided to work on this project because we found AI interesting and we thought that this project would lend itself to efficient delegation of work and collaboration amongst team members.
- Our project will implement at least two different machine learning models, one for classification and one for regression
- We will also support model customization and hyperparameter customization for both models
- We aim to have the first model type without customization completed by the first checkpoint, the second model type without customization completed by the second checkpoint, and customization for both models (e.g., regularization, etc.) completed by the final deadline
- One challenge we might run into is implementing all of the math behind machine learning, as we may have to learn some math concepts on the fly
- Another challenge we might face is coding all of our algorithms efficiently so that model training and evaluation does not take too much memory or time
- Complex syntax in rust might make it difficult for us to create our project
The python library scikit-learn provided us with the general inspiration for this project.