Machine Learning for Healthcare

Machine Learning for Healthcare

Some cases, opportunities, and challenges in applying machine learning to healthcare problems.


Ali Akbar S.

April 27, 2019


  1. Machine Learning for Healthcare Ali Akbar Septiandri @aliakbars

  2. • What is machine learning (ML)? • Opportunities in machine

    learning for healthcare (ML4HC) • Challenges in ML4HC • Future of ML4HC Outline
  3. What is machine learning?

  4. How to identify cats or dogs in an image?

  5. Identifying Cats or Dogs Model • Logit model • SVM

    Image Processing • Edge detection • Texture analyser • Color histogram Feature Extraction • Eye position • Eye colour • Nose colour • Fur type • Leg counts
  6. y = σ(β 0 + β 1 x 1 +

    β 2 x 2 + β 3 x 3 ) Logit model from defined features
  7. “Fundamentally, machine learning involves building mathematical models to help understand

    data.” - Jake VanderPlas
  8. …but it might not even work!

  9. Enter: Deep Learning

  10. (Artificial) Neural Networks ~ Deep Learning

  11. It’s that simple*! Output Feature extraction + model training Input

  12. “No free lunch”

  13. Opportunities

  14. Verbal screening using mobile app

  15. Stanford ML Group

  16. None
  17. “ could diagnose 14 diseases with 80% accuracy–in other words,

    about as well as a real radiologist.”
  18. Deep Neural Networks Improve Radiologists’ Performance in Breast Cancer Screening

    (Wu et al., 2019)
  19. None
  20. None
  21. None
  22. Making out of focus microscopy images in focus again

  23. Early Warning System System for Electronic Notification and Documentation (SEND)

    by Sensyne Health
  24. Natural Language Processing (NLP) for clinical notes

  25. DeepMind’s AlphaFold

  26. “...diagnosing and treating diseases believed to be caused by misfolded

    proteins, such as Alzheimer’s, Parkinson’s, Huntington’s and cystic fibrosis.” (Evans et al., 2018)
  27. Challenges

  28. Accuracy

  29. This is an image of an apical pneumothorax!

  30. “Correlation does not imply causation” | Image from xkcd

  31. Understanding Causality Graphical presentation of confounding in directed acyclic graphs

    (Suttorp et al., 2014)
  32. Missing Data

  33. Adversarial Attacks (Finlayson et al., 2019)

  34. Bias in ML “...a project to look for skin cancer

    in photographs. It turns out that dermatologists often put rulers in photos of skin cancer, for scale, but that the example photos of healthy skin do not contain rulers. To the system, the rulers (or rather, the pixels that we see as a ruler) were just differences between the example sets, and sometimes more prominent than the small blotches on the skin. So, the system that was built to detect skin cancer was, sometimes, detecting rulers instead.” (Evans, 2019)
  35. Future of ML4HC

  36. Optimal treatment strategies, e.g. for sepsis (Komorowski et al., 2018)

  37. Generating Images from Brain Signals

  38. None
  39. Where are we now?

  40. Thank you. @__aliakbars__