Natural Language Processing

With applications in Medical Diagnosis from Pathology Reports

Author

Emil Sargsyan

Preface

This book is a community-college-friendly companion to AI Campus @ Cedars-Sinai Project 7: NLP in Cancer Pathology Reports.

It accompanies the public notebook repository at https://github.com/emilsar/NLP-TCGA, walking students from text classification basics through transformer fine-tuning and LLM prompting on a real clinical-NLP corpus.

What you will learn

How to turn 9,000+ scanned pathology reports into a labeled NLP dataset.
Classical baselines: bag-of-words, TF-IDF, logistic regression, random forest.
Transformer-based classification with Clinical-BigBird.
LLM prompting for structured field extraction.
How to measure progress honestly with held-out splits and per-class metrics.

Reference projects

Teaching repo: guilopgar/AI-Campus-Project-7-NLP — pre-cleaned data and notebooks.
Research repo: tatonetti-lab/tnm-stage-classifier — published models (BBTEN) with external validation.

Prerequisites

Comfort with Python, pandas, and Jupyter. No prior NLP or ML background assumed.