AIMar 27, 2013
Learning Link-Probabilities in Causal Trees
arXiv:1304.3103v13 citations
Originality Synthesis-oriented
AI Analysis
This addresses a specific challenge in causal inference for researchers dealing with hidden variables, but it appears incremental in nature.
The paper tackles the problem of estimating link probabilities in causal trees with hidden internal nodes using only leaf measurements, presenting an algorithm that is incremental, local, efficient, and robust to measurement imprecisions.
A learning algorithm is presented which given the structure of a causal tree, will estimate its link probabilities by sequential measurements on the leaves only. Internal nodes of the tree represent conceptual (hidden) variables inaccessible to observation. The method described is incremental, local, efficient, and remains robust to measurement imprecisions.