Component Caching in Hybrid Domains with Piecewise Polynomial Densities

Authors: Vaishak Belle, Guy Van den Broeck, Andrea Passerini

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Evaluations In this section, we demonstrate that the approach introduced here performs significantly better than known techniques for model counting of arithmetic constraints. Our experiments demonstrate a large gap in performance when compared to existing approaches based on a variety of block-clause strategies.
Researcher Affiliation Academia Vaishak Belle KU Leuven, Belgium vaishak@cs.kuleuven.be Guy Van den Broeck University of California, Los Angeles guyvdb@cs.ucla.edu Andrea Passerini University of Trento, Italy passerini@disi.unitn.it
Pseudocode Yes Algorithm 1 #DPLLCache(Δ, w) returns WMC(Δ, w)
Open Source Code No The paper states: 'Our techniques are built on the competitive WMC solver cachet (Sang, Beame, and Kautz 2005). For integrating in fixed dimensions efficiently, we use Latt E v1.6.5' and 'BC is implemented using Math SAT v5.' but does not provide a link to the authors' own implementation code.
Open Datasets No The paper states: 'We consider such hybrid versions of alarm, water, insurance, and child,' which were constructed by randomly modeling nodes as continuous variables from standard Bayesian networks, but no specific access information (link, DOI, repository) for these hybrid datasets is provided.
Dataset Splits No The paper does not explicitly provide specific dataset split information (e.g., percentages, sample counts, or cross-validation details) for training, validation, or testing.
Hardware Specification Yes All experiments were run using a 2.83 GHz Intel Core 2 Quad processor and 8GB RAM.
Software Dependencies Yes Our techniques are built on the competitive WMC solver cachet (Sang, Beame, and Kautz 2005). For integrating in fixed dimensions efficiently, we use Latt E v1.6.5. ... BC is implemented using Math SAT v5.
Experiment Setup Yes Given a timeout of 12 hours... We first find a model M for the input sentence, which is a complete truth assignment to all the propositions. Suppose there are n propositions, and we would like to provide k propositions as evidence. Then the SMT and CNF files are modified to assert the truth values of k propositions as suggested by M.