<ul><li>A new research paper introduces results for bounding the value of a functional of the target distribution, such as the generalization error of a classifier, given data from source domains and assumptions about the data generating mechanisms.</li><li>The paper builds on the theory of partial identification and transportability to provide the first general estimation technique for transportability problems.</li><li>The authors adapt existing parameterization schemes, such as Neural Causal Models, to encode the necessary structural constraints for cross-population inference.</li><li>The paper also proposes a gradient-based optimization scheme to make scalable inferences in practice, and the results are supported by experiments.</li></ul>

Partial Transportability for Domain Generalization

Discover more