Mathematics · advanced
Maximum Likelihood Estimation
Choosing parameters that make the observed data most probable; log-likelihood; connection to cross-entropy loss.
Mental model
MLE picks the parameter values that maximize the probability of what you actually saw. Log-likelihood turns products into sums — easier to optimize. Minimizing cross-entropy on classification data is MLE under a Bernoulli/categorical model.
How to study Maximum Likelihood Estimation
Begin by restating the mental model in your own words, then connect it to a concrete system you have built or operated. Name the mechanism, the constraint it addresses, and the trade-off it introduces. Use StatQuest — Maximum Likelihood to check details, but close the source before writing your explanation. Retrieval is the learning step; rereading is only preparation.
Next, compare Maximum Likelihood Estimation with Linear Regression, Gradient Descent, Softmax & Cross-Entropy, Information & Entropy. Ask what changes in correctness, latency, resource use, operability, and failure recovery. Complete MLE for Bernoulli trials and preserve the command, input, output, and one failed attempt as evidence. Finish by explaining the idea without jargon to someone who has not studied the track.
Proof of understanding
- Explain the mechanism from first principles and identify the state it reads or changes.
- Give one situation where the concept is the right choice and one where it is not.
- Predict a realistic failure mode before running the drill, then compare the prediction with evidence.
- Connect the result to a roadmap or build artifact instead of treating the concept as isolated trivia.
Where it matters
Training classifiers, fitting regression, EM algorithms, calibration.
Common mistakes
- Maximizing likelihood on training data with no regularization (overfitting)
- Forgetting that MLE is a point estimate, not uncertainty
- Mixing up likelihood P(data|θ) with probability of θ
Learn from primary sources
- StatQuest — Maximum Likelihood (video)
Practice and explain it back
MLE for Bernoulli trials
n=100 Bernoulli trials, 55 successes. Write the likelihood L(p)=p^55(1−p)^45, take log, differentiate, and solve for p̂. Why is this the same as the sample proportion?
Expected evidence: p̂=55/100=0.55. Log-likelihood derivative gives successes/n — MLE equals the intuitive estimator here.
Open the interactive drill →Review prompts
- How does MLE relate to cross-entropy loss?
- Likelihood vs posterior?
Build evidence
Use a roadmap capstone to turn this concept into working evidence.