coursera-ai4med-course3

coursera-ai4med-course3

TODOs

• Fix links
• Ensure large files are available for notebooks

Week 1

Key Concepts

• Analyze data from a randomized control trial
• Interpreting Multivariate Models
• Evaluating Treatment Effect Models
• Interpreting ML models for Treatment Effect Estimation

Notebooks

• week1a
  • i.e., pandas refresher, slicing, read/update rows, read/update columns
• week1b
  • i.e., train/test split, simple model learn/fit/evaluate
  • i.e., dict args as **kwargs
  • i.e., itertools.product() to derive permutations
  • i.e., pass *args list
• week1c
  • i.e., logistic regression

Assignment

• assignment1

References

• RCT
• Levamisole and fluororacil background: https://www.nejm.org/doi/full/10.1056/NEJM199002083220602
• Data sourced from here: https://www.rdocumentation.org/packages/survival/versions/3.1-8/topics/colon
• C-statistic for benefit: https://www.ncbi.nlm.nih.gov/pubmed/29132832
• T-learner: https://arxiv.org/pdf/1706.03461.pdf

Week 2

Key Concepts

• Extracting disease labels from clinical reports
• Question Answering with BERT

Notebooks

• week2a

References

• BERT paper (Google, 2018)
• BERT Github
• NegBio
• [Grad cam])https://arxiv.org/pdf/1610.02391.pdf)
• Random forests + permutation importance: https://www.stat.berkeley.edu/~breiman/randomforest2001.pdf (R45f14345c000-1 Breiman, “Random Forests”, Machine Learning, 45(1), 5-32, 2001.)
• Shapley importance: https://www.nature.com/articles/s42256-019-0138-9

Week 3

Key Concepts

• Interpreting Deep Learning Models
• Feature Importance in Machine Learning

Notebooks

• lesson3a
  • i.e., aggregate and group CCS codes; learn mappings

Lesson 4

Notebooks

• lesson4a
  • i.e., transform line to encounter
  • i.e., pandas groupby()

All References

• Labeling methods and dataset
• Huggingface transformers library
• BERT paper
• Question answering data set (used for example)
• Clinical note example for question answering