Bayesian inference

Ever wondered how doctors update diagnoses as new symptoms arise?

Imagine you're feeling sick and first see a doctor. They ask about your symptoms and make an initial diagnosis. As time goes on and you develop new symptoms, the doctor revises their diagnosis based on all the new information.

The doctor starts with an initial guess (prior) about what's wrong based on your first symptoms. As you get sicker and new symptoms appear, they update their guess (posterior) using all the new information. This process of updating is called Bayesian inference.

Example

If the doctor initially thinks it's a common cold (prior probability), but then you start experiencing unusual symptoms like a rash (new evidence), they'll update their diagnosis to consider other illnesses (posterior probability).

Remember this

Doctors use Bayesian inference to refine diagnoses as new symptoms emerge, ensuring they consider all available information.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

Conjugate prior

Conjugate priors simplify Bayesian updating

the Dirichlet distribution does: distribution over probability simplices

How do we predict the likelihood of various outcomes in uncertain situations?

Bayes' theorem

Bayes' theorem formula: P(A|B) = [P(B|A) * P(A)] / P(B)

Fisher information

Fisher information measures information about unknown parameters

the do-calculus does: computes interventional probabilities from observational data

Do-calculus computes interventional probabilities from observational data

Markov chain Monte Carlo

MCMC samples from complex posterior distributions

Swipe through 100 ML concepts daily

Open Pocket Polymath