solulubeta
Sign in
Cheat Sheets / System Design and ML
04.02 · Cheat sheet

Machine Learning

Types of learning, bias vs variance, evaluation metrics, the main algorithms, gradient descent, neural networks, data preparation and LLM basics on one page.

3 min Cheat sheet Pro

The ML ideas and formulas that come up in data and ML interviews.

Types of learning

  • Supervised: labelled data. Regression (a number) or classification (a class).
  • Unsupervised: no labels. Clustering, dimensionality reduction.
  • Reinforcement: an agent learns from rewards.
  • Self-supervised: labels come from the data itself (e.g. predict the next word).

Bias vs variance

UnderfittingOverfitting
ProblemHigh biasHigh variance
Train errorHighLow
Test errorHighHigh
FixBigger model, more featuresMore data, regularisation, simpler model
  • L1 pushes weights to exactly 0 (feature selection). L2 keeps weights small.
  • Also: dropout, early stopping, k-fold cross-validation.
solulu Pro

Continue this cheat sheet with Pro

Inside: 9 sections.

  • Types of learning
  • Bias vs variance
  • Classification metrics
  • Regression metrics
  • Main algorithms
  • Gradient descent
  • +3 more
Sign in to continue See plans
Lifetime ₹1,999 ₹999 · Launch offer
Mark this as done when you are finished.

Item 12 of 12 · updated 2026-10-13