AI/ML · Research
M.A. thesis · York University
ASC-PIE
Every PII dataset spoke its own language. I built a shared one, then put seven models through the same exam.
I developed this PII-aware named-entity recognition corpus and evaluation framework for my M.A. thesis, harmonizing four public corpora plus my own synthetic augmentation under a shared 19-type schema.
- 333,109examples
- 99.1%best strict F1
My roleDesigned and built the research pipeline, schema and evaluation framework.
- Python
- PyTorch
- Hugging Face Transformers
- Hugging Face Datasets
- scikit-learn