liveBook · Manning

Probability of improvement

Probability of improvement is an acquisition function that estimates the probability a candidate will exceed a specified incumbent objective value or improvement threshold.

How 2 Manning books explain it

Bayesian Optimization in Action

Probability of Improvement is a Bayesian optimization acquisition function that assigns each candidate the probability that its objective value will exceed a specified incumbent or improvement threshold.

How it is used
Selects the next expensive black-box evaluation by prioritizing candidates likely to beat the incumbent.
How deep it goes
Derives PoI as a normal-distribution tail area, implements it in BoTorch, and examines its iterative behavior and epsilon modification.
Examples
Uses validation accuracy, alloy stability, the one-dimensional Forrester function, and two-dimensional SVM hyperparameter tuning.
What you can do after
Apply it via BoTorch to optimize a PoI acquisition score and adjust epsilon when exploitation stalls progress.

Read how Bayesian Optimization in Action explains probability of improvement →

Build AI Drug Discovery Pipelines

Probability of improvement is an acquisition criterion that estimates the probability that a candidate will outperform the best observed result.

How it is used
Prioritizes molecules for expensive binding-affinity evaluation within an active-learning virtual-screening pipeline.
How deep it goes
Defines PI and its selection role, but defers the implementation details to a separate notebook.
Examples
Uses molecular pools, binding-affinity predictions, docking, and a VEGFR-2 experiment recovering 73 of the top 100 molecules.
What you can do after
Apply PI to prioritize the next molecules for expensive evaluation in an active-learning screening loop.

Read how Build AI Drug Discovery Pipelines explains probability of improvement →

What to know first

Related concepts

Questions about probability of improvement

What is probability of improvement?

Probability of improvement is an acquisition function that estimates the probability a candidate will exceed a specified incumbent objective value or improvement threshold.

Which Manning books explain probability of improvement?

Bayesian Optimization in Action and Build AI Drug Discovery Pipelines.

What should I know before learning probability of improvement?

Manning books that teach probability of improvement usually introduce Gaussian process and standard deviation first.

What is probability of improvement a kind of?

Probability of improvement is a kind of Bayesian optimization. Bayesian optimization is a sequential method for optimizing an expensive or unknown objective by using a probabilistic model to guide evaluations.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.