Compiling Bayesian Network Classifiers into Decision Graphs

Authors: Andy Shih, Arthur Choi, Adnan Darwiche7966-7974

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results relating to scalability are then given in Section 6, followed by a case study in Section 7.
Researcher Affiliation Academia Andy Shih, Arthur Choi, Adnan Darwiche Computer Science Department University of California, Los Angeles {andyshih,aychoi,darwiche}@cs.ucla.edu
Pseudocode Yes Algorithm 1 compile-classifier(B, π) ... Algorithm 2 compile-subclassifier(B, u, π, k) ... Algorithm 3 block-order(B, X)
Open Source Code No The paper mentions using the 'Sam IAm library' but does not provide access to the authors' own implementation code for their proposed method.
Open Datasets Yes The win95pts network is used to diagnose why a printing job has failed (Breese and Heckerman 1996). ... Next, we consider the Andes network, which models students problem-solving skills in physics (Gertner, Conati, and Van Lehn 1998).
Dataset Splits No The paper does not provide specific dataset split information (e.g., percentages, sample counts) for training, validation, and test sets.
Hardware Specification No The paper does not provide specific hardware details (e.g., CPU/GPU models, memory) used for running its experiments.
Software Dependencies No Inference calls were performed using the Sam IAm library.2 ... 2Available at http://reasoning.cs.ucla.edu/samiam/ (No version number mentioned for the library.)
Experiment Setup Yes We assume a threshold of 1/2. ... The ODD size and compilation time are also significantly affected by the threshold of the classifier. A heavily biased threshold can lead to a very small ODD and a short compilation time, while a balanced threshold generally leads to larger ODDs.