Ever wondered how scientists predict outcomes with uncertainty?
Ever wondered how scientists predict outcomes with uncertainty?
Imagine you're trying to guess the weather for a picnic. You know it's usually sunny, but sometimes it rains unexpectedly.
Think of the weather like a bag of dice, where each die represents a different weather pattern. The Dirichlet process helps us understand the mix of these patterns and predict the weather more accurately.
Example
If you've guessed sunny weather correctly 70% of the time in the past, the Dirichlet process helps you figure out the chances of rain or sunshine for the next picnic.
Remember this
The Dirichlet process is like a smart guesser that learns from past weather patterns to predict future ones.
Text adapted from Wikipedia, licensed under CC BY-SA 4.0.
the Dirichlet distribution does: distribution over probability simplices
How do we predict the likelihood of various outcomes in uncertain situations?
Likelihood function
Why can't we just guess probabilities correctly all the time?
Entropy H = -Σ p(x) log₂ p(x) measures average surprise in bits
How do we measure uncertainty in everyday decisions?
log-probabilities are used instead of probabilities: avoids numerical underflow
Why can't we just add up tiny chances over time?
Markov chain Monte Carlo
MCMC samples from complex posterior distributions
Exponential distribution
Why do we sometimes need to predict rare events?
Swipe through more Machine Learning concepts
Open Pocket Polymath