Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Searching for the M Best Solutions in Graphical Models

Authors: Natalia Flerova, Radu Marinescu, Rina Dechter

JAIR 2016 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments confirm theory that the best-first approach is largely superior when memory is available, but depth-first branch and bound is more robust. We also show that our algorithms are competitive with related schemes recently developed for the m-best task.
Researcher Affiliation Collaboration Natalia Flerova EMAIL University of California, Irvine Irvine, CA 92697, USA Radu Marinescu EMAIL IBM Research Ireland Rina Dechter EMAIL University of California, Irvine Irvine, CA 92697, USA
Pseudocode Yes Algorithm 1: AOBF exploring the AND/OR search tree (Marinescu & Dechter, 2009b)... Algorithm 2: AOBB exploring the AND/OR search tree (Marinescu & Dechter, 2009b)... Algorithm 3: Recursive computation of the heuristic evaluation function... Algorithm 4: Mini-Bucket Elimination... Algorithm 5: m-A* exploring a graph, assuming consistent heuristic... Algorithm 6: m-BB exploring a graph, assuming a consistent heuristic... Algorithm 7: m-AOBF exploring an AND/OR search tree... Algorithm 8: m-AOBB exploring an AND/OR search tree... Algorithm 9: Combining the sets of costs and partial solution trees... Algorithm 10: Merging the sets of costs and partial solution trees
Open Source Code No The paper does not provide an explicit statement or link for the open-sourcing of the authors' implementation for the methodology described. It mentions that implementations of competing schemes were provided by Dhruv Batra, but not their own code.
Open Datasets Yes The pedigrees benchmark ( pedigree* ) was used in the UAI 2008 competition.3 They arise from the domain of genetic linkage analysis and are associated with the task of haplotyping. 3. http://graphmod.ics.uci.edu/group/Repository In each of the binary grid networks ( 50-* , 75-* and 90-* )4 the nodes corresponding to binary variables are arranged in an N by N square and the functions are defined over pairs of variables and are generated uniformly randomly. 4. http://graphmod.ics.uci.edu/repos/mpe/grids/
Dataset Splits No The paper describes the benchmarks used and how instances were generated for some, but does not provide specific train/test/validation splits (e.g., percentages or counts) or a methodology for creating such splits that would be needed for reproduction.
Hardware Specification Yes All algorithms were implemented in C++ (32-bit) and the experiments were run on a 2.6GHz quad-core processor. The memory limit was set for 4 GB per problem, the time limit to 3 hours.
Software Dependencies No The paper states that algorithms were "implemented in C++ (32-bit)" but does not provide specific version numbers for the C++ compiler or any libraries used, which are essential for reproducibility.
Experiment Setup Yes We used 10 i-bounds, ranging from 2 to 22. However, for some hard problems computing the mini-bucket heuristic with the larger i-bounds proved infeasible, so the actual range of i-bounds varies among the benchmarks and among instances within a benchmark. All algorithms were restricted to a static variable ordering computed using a min-fill heuristic (Kjærulff, 1990). Both AND/OR schemes used the same pseudo tree.