- Computer Science
- Published in ICML 2019

# Conditional Independence in Testing Bayesian Networks

@inproceedings{Shen2019ConditionalII, title={Conditional Independence in Testing Bayesian Networks}, author={Yujia Shen and Haiying Huang and Arthur Choi and Adnan Darwiche}, booktitle={ICML}, year={2019} }

Proposition 2 Suppose dsepG(X,Z,Y) holds but dsepG(U \ Z,Z,Y) does not hold. We have U \Y = ; by dsepG(X,Z,Y). Moreover, there must exist a path connecting some U 2 U \ Z and Y that is not blocked by Z. If X is on this path, then we have a path connecting X and Y that is not blocked by Z. Otherwise, augmenting this path with the edge X U leads to a path with the same properties. Either case contradicts dsepG(X,Z,Y). ⇤

#### References

##### Publications referenced by this paper.

SHOWING 1-10 OF 12 REFERENCES

## On the Relative Expressiveness of Bayesian and Neural Networks

VIEW 6 EXCERPTS

HIGHLY INFLUENTIAL

## On the relative expressiveness of Bayesian and neural networks

## On Relaxing Determinism in Arithmetic Circuits

VIEW 2 EXCERPTS

## On the Expressive Power of Deep Neural Networks

VIEW 1 EXCERPT

## Comment: Understanding Simpson’s Paradox

VIEW 1 EXCERPT

## Approximation by Superpositions of a Sigmoidal Function

VIEW 1 EXCERPT

## A Differential Approach to Inference in Bayesian Networks

VIEW 3 EXCERPTS