Project Instructions:
Navigate to this notebook and select "run before" from the runtime menu while in the last cell. Website can be accessed through the link created in the output of the previous cell.
From Syllabus:
This course has a specific interest in using AI systems in software products. The course will look at ways to add an AI component to an existing web-based system to enhance the system’s capabilities. The course will take a human-in-the-loop approach, augmenting a human user instead of attempting to replace them. Because of this, issues with explainability must be considered.
Specific topics we will take up are:
- Managing the steps (AKA the workflow) of data science that allow us to transform raw data into production-ready form. The jargony term is MLOps.
- Managing the model-exploration phase to get to a production-ready form. We will study individual models (e.g., Regression, XGBoost, KNN, Neural Nets) and ensembles.
- Putting data and models together in a website that augments a user with AI assistance.
- Giving users the ability to judge AI results by providing effective explanations.
Taught by Prof. Stephen Fickas at University of Oregon.