Skip to content

Learn track — zero to research-ready

Assumes no AI/ML and no medical background. Read in order; each lesson lists what to read next. Technical detail lives in the research section; this track explains ideas.

Map to research phases in ../research/plan.md:

  • Phase 1 (literature review): lessons 01–05, then 11
  • Phase 2 (data + baselines): lessons 06, 08
  • Phase 3 (Approach A): lessons 07, 09, 10
  • Phase 4–5 (comparison, write-up): revisit 04, 10, 11

Lessons:

  • 01-ecg-heart-basics.md — what the heart does, what ECG measures, PQRST
  • 02-arrhythmia-and-aami-classes.md — what arrhythmia is, N/V/S/F/Q in plain words
  • 03-ml-classification-basics.md — labels, training, test splits, leakage
  • 04-metrics.md — precision, recall, F1, why accuracy lies
  • 05-class-imbalance.md — rare beats, weighting, sampling
  • 06-ecg-signal-processing.md — sampling, baseline wander, windows, normalization
  • 07-scalograms.md — wavelets, CWT heatmaps, reading a scalogram
  • 08-neural-nets-cnn-lstm.md — neurons, CNNs, LSTMs, what baselines are
  • 09-ssm-mamba.md — state-space models and Mamba intuition, no heavy math
  • 10-fusion-loss-thresholds.md — logits, sigmoid, BCE, fusion, thresholds
  • 11-reading-papers.md — how to read ECG papers, splits, leakage red flags