In Fall 2023, UVicAI hosted a series of educational workshops to get our members introduced to TensorFlow, a popular machine learning framework and an entry point for the core ideas around this field. This series spanned multiple weeks and touched three broad topics: image classification, time series, and natural language. The selection of these topics was informed by the TensorFlow certification exam.
Corresponding slides can be found linked in UVicAI's google drive.
The Timeseries notebook references "airport data", which is available through the Google Drive link above.
The "sunspots" dataset is a well-utilized timeseries dataset, but this version has the final year clipped. This is for a timeseries regression task where members didn't have access to the test set but had to determine how to validate their model on just the existing data.
The "crowd trajectory" dataset is from Ohio State University (Paper - Arxiv). It was not used in this series, but was discussed as an interesting application related to sequential data.