And here’s a really fascinating paper by some of my colleagues that tries to model causal interventions that go beyond Pearl’s do-operator, by formalizing causal interventions as (probabilistic) program transformations:
Bayesian causal inference via probabilistic program synthesis Sam Witty, Alexander Lew, David Jensen, Vikash Mansinghka https://arxiv.org/abs/1910.14124
Causal inference can be formalized as Bayesian inference that combines a prior distribution over causal models and likelihoods that account for both observations and interventions. We show that it is possible to implement this approach using a sufficiently expressive probabilistic programming language. Priors are represented using probabilistic programs that generate source code in a domain specific language. Interventions are represented using probabilistic programs that edit this source code to modify the original generative process. This approach makes it straightforward to incorporate data from atomic interventions, as well as shift interventions, variance-scaling interventions, and other interventions that modify causal structure. This approach also enables the use of general-purpose inference machinery for probabilistic programs to infer probable causal structures and parameters from data. This abstract describes a prototype of this approach in the Gen probabilistic programming language.
What useful problems do PPLs solve? Ideally some applications that are interesting for us non-corporate people. Can it be used for medical statistics (e.g., in nutrition)? (Any examples?) Is the reason it is not used the illiteracy of the scientists, or are the mainstream methods better?
PPLs are a tool to bring complicated statistical modeling to the masses. Computers are capable of doing much more advanced statistical modeling than appears in every non-statistics paper, but most people don’t have the expertise to build them. PPLs allow you to write complicated statistical models and then evaluate them with state-of-the-art methods without having to build everything from scratch.
Belatedly seeing this post, but I wanted to note that probabilistic programming languages (PPLs) are centered around this basic idea! Some useful links and introductions to PPLs as a whole:
- Probabilistic models of cognition (web book)
- WebPPL
- An introduction to models in Pyro
- Introduction to Modeling in Gen
And here’s a really fascinating paper by some of my colleagues that tries to model causal interventions that go beyond Pearl’s do-operator, by formalizing causal interventions as (probabilistic) program transformations:
What useful problems do PPLs solve? Ideally some applications that are interesting for us non-corporate people. Can it be used for medical statistics (e.g., in nutrition)? (Any examples?) Is the reason it is not used the illiteracy of the scientists, or are the mainstream methods better?
PPLs are a tool to bring complicated statistical modeling to the masses. Computers are capable of doing much more advanced statistical modeling than appears in every non-statistics paper, but most people don’t have the expertise to build them. PPLs allow you to write complicated statistical models and then evaluate them with state-of-the-art methods without having to build everything from scratch.