Machine Learning in One Sitting

The whole field compresses to a loop: guess, measure error, adjust, repeat. Everything else is engineering around that loop.

Pooja Thirupuranthakam · May 15, 2025 2 min read 6

Strip machine learning to its core and you get one loop:

  1. Make a model with adjustable parameters.
  2. Make a prediction.
  3. Measure how wrong it was (the loss).
  4. Nudge the parameters to be slightly less wrong.
  5. Repeat on lots of data.

That's it. Linear regression, random forests, and trillion-parameter language models are all this loop with different choices for "model" and "nudge".

The vocabulary you actually need

  • Features: the inputs. Labels: the correct answers (for supervised learning).
  • Training set / validation set / test set: what you learn from, what you tune on, what you never touch until the end. Mixing these up is the most common beginner mistake and it makes your results fiction.
  • Overfitting: the model memorized the training set and fails on new data. Symptom: training loss keeps dropping while validation loss rises.
  • Regularization: any trick that fights overfitting (penalties on parameter size, dropout, early stopping, more data).
  • Generalization: performing well on data you've never seen. The only thing that matters.

Three families

Family Learns from Example
Supervised (input, label) pairs spam detection, image classification
Unsupervised inputs only clustering customers, compression
Reinforcement rewards from an environment game playing, robot control, RLHF

Language models blur these: pretraining is self-supervised (predict the next token; the label is the data itself), then fine-tuning and RL shape behavior.

Bias and variance, in one sentence each

High bias: the model is too simple to capture the pattern (a line through a curve). High variance: the model is so flexible it fits noise. You trade one for the other; more data lets you afford more variance.

The practical workflow

look at the data (really look)  →  baseline (mean, majority class, logistic regression)
→ better model  →  measure on validation  →  error analysis: what does it get wrong?
→ fix data/features/model  →  repeat  →  test set once, at the end

Most wins come from the "look at the data" and "error analysis" steps, not the model. A senior ML engineer spends most of their time there.

What to learn next

Gradient descent (how the nudge works), then deep learning (what happens when the model is a stack of layers), then evaluation (how not to fool yourself). Skip the zoo of classical algorithms until you need one.

Written by

Pooja Thirupuranthakam

Senior in Artificial Intelligence at Purdue. I write about how modern AI systems actually get built.

Discussion (0)

  • No comments yet. Start the thread.

Log in to join the discussion.