Unity Smoothing for Handling Inconsistent Evidence in Bayesian Networks and Unity Propagation for Faster Inference
This work addresses efficiency and accuracy issues in Bayesian network inference for applications relying on sparse data, though it is incremental as it builds on existing methods like Laplace smoothing and the junction tree algorithm.
The authors tackled the problem of inconsistent evidence in Bayesian networks by proposing Unity Smoothing (US), which achieved prediction accuracy comparable to Laplace smoothing while reducing memory usage in sparse data structures, and introduced Unity Propagation (UP) to speed up inference by avoiding redundant calculations in the junction tree algorithm.
We propose Unity Smoothing (US) for handling inconsistencies between a Bayesian network model and new unseen observations. We show that prediction accuracy, using the junction tree algorithm with US is comparable to that of Laplace smoothing. Moreover, in applications were sparsity of the data structures is utilized, US outperforms Laplace smoothing in terms of memory usage. Furthermore, we detail how to avoid redundant calculations that must otherwise be performed during the message passing scheme in the junction tree algorithm which we refer to as Unity Propagation (UP). Experimental results shows that it is always faster to exploit UP on top of the Lauritzen-Spigelhalter message passing scheme for the junction tree algorithm.