score matching does: learns the gradient of the log-density without normalizing

Propensity score matching reduces bias in treatment effect estimates

In randomized experiments, randomization ensures unbiased estimation of treatment effects by balancing treatment groups on average for each covariate. However, PSM is used when randomization is not possible, allowing for the estimation of treatment effects while accounting for confounding variables.

Example

In a study comparing the effects of a new drug to a placebo, researchers used PSM to match patients based on age, gender, and other health indicators. This helped to ensure that the treatment groups were comparable, reducing bias in the estimated effect of the drug.

Remember this

Understanding PSM is crucial for accurately estimating treatment effects in observational studies where randomization is not possible.

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

Related concepts

denoising score matching does: learns to denoise, which equals learning the score

Propensity score matching (PSM) reduces bias in treatment effect estimates

the back-door criterion identifies: sufficient adjustment sets for causal estimation

Causal models use formal notation like DAGs for causal inference

to standardize: when you need zero mean and unit variance for gradient-based optimization

Why do we need to make data uniform before training a model?

Regression discontinuity design

RDD uses a sharp threshold for treatment assignment

Reinforcement learning from human feedback

RLHF optimizes a reward model trained on human preference pairs

Race and intelligence

IQ test performance differences between racial groups have decreased over time

Swipe through 100 ML concepts daily

Open Pocket Polymath