Anytime Anyspace AND/OR Search for Bounding the Partition Function
Authors: Qi Lou, Rina Dechter, Alexander Ihler
AAAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Empirical results demonstrate that our approach with heuristics extracted from weighted mini-bucket (Liu and Ihler 2011) is almost always superior to the baselines on various benchmark-memory settings. and Empirical Evaluation To evaluate the effectiveness of our algorithm in Alg. 2, we compare it to a number of existing methods on four benchmarks and three different memory budgets. |
| Researcher Affiliation | Academia | Qi Lou University of California, Irvine Irvine, CA 92697, USA qlou@ics.uci.edu Rina Dechter University of California, Irvine Irvine, CA 92697, USA dechter@ics.uci.edu Alexander Ihler University of California, Irvine Irvine, CA 92697, USA ihler@ics.uci.edu |
| Pseudocode | Yes | Algorithm 1 Best-first search for anytime sum bounds. and Algorithm 2 Memory-limited BFS for anytime sum bounds. |
| Open Source Code | No | The paper thanks other researchers for sharing code/data but does not provide an explicit link or statement of open-source release for the code of the methodology described in this paper. |
| Open Datasets | No | We evaluated performance on several benchmark instance sets: CPD, a set of computational protein design problems from Viricel et al. (2016); PIC 11, a benchmark subset of 23 instances selected by Viricel et al. (2016) from the 2012 UAI competition; BN, a set of Bayesian networks from the 2006 competition1; and Protein, made from the small protein side-chains of Yanover and Weiss (2002). The paper provides citations to papers that might describe the datasets, but not direct access links for the datasets themselves. |
| Dataset Splits | No | The paper mentions evaluating performance on benchmark instance sets and discusses memory and time budgets, but it does not provide specific details on training, validation, or test dataset splits or percentages. |
| Hardware Specification | No | The paper specifies memory budgets (1GB, 4GB, and 16GB) but does not provide specific details on the CPU, GPU, or other hardware components used for running the experiments. |
| Software Dependencies | No | The paper states that 'Implementations of all methods are in C/C++' but does not specify any software names with version numbers, such as compilers or specific libraries. |
| Experiment Setup | Yes | For a given memory budget, we first compute the largest ibound that fits the memory budget, and then use the remaining memory for search. and We adopted the parameter setup suggested by the authors: upper bound set to Ub1, enforcing VAC at the root, and using EDAC during the rest of the search. |