Latent Tree Models and Approximate Inference in Bayesian Networks
This work addresses computational efficiency for users of Bayesian networks, but it is incremental as it builds on existing latent tree model techniques.
The authors tackled the problem of slow inference in Bayesian networks by proposing a method that uses latent tree models for approximate inference, achieving good accuracy with linear-time online computation.
We propose a novel method for approximate inference in Bayesian networks (BNs). The idea is to sample data from a BN, learn a latent tree model (LTM) from the data offline, and when online, make inference with the LTM instead of the original BN. Because LTMs are tree-structured, inference takes linear time. In the meantime, they can represent complex relationship among leaf nodes and hence the approximation accuracy is often good. Empirical evidence shows that our method can achieve good approximation accuracy at low online computational cost.