A Novel Approach for Constrained Optimization in Graphical Models

Authors: Sara Rouhani, Tahrima Rahman, Vibhav Gogate

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments show that our algorithms are superior to the following approach: encode the problem as a mixed integer linear program (MILP) and solve the latter using a state-of-the-art MILP solver such as SCIP.
Researcher Affiliation Academia Sara Rouhani, Tahrima Rahman and Vibhav Gogate The University of Texas at Dallas {sara.rouhani,tahrima.rahman,vibhav.gogate}@utdallas.edu
Pseudocode Yes Algorithm 1 ANYTIME-CMPE (M1, M2, q, k)
Open Source Code No The paper mentions using SCIP [13], a state-of-the-art open source MILP solver, but does not provide specific access to the authors' own implementation code for the methodology described.
Open Datasets Yes We experimented with the following benchmark graphical models, available from the UAI 2010 and 2014 competitions [11, 16]
Dataset Splits No The paper mentions using benchmark graphical models but does not provide specific information about training, validation, or test dataset splits.
Hardware Specification No No specific hardware details (such as GPU/CPU models or memory specifications) used for running the experiments are mentioned in the paper.
Software Dependencies Yes We compared the performance of Algorithm ANYTIME-CMPE with SCIP [13], a state-of-the-art open source mixed integer linear programming (MILP) solver. ... The SCIP Optimization Suite 7.0.
Experiment Setup Yes We experimented with the following five values of k: {1, 3, 5, 7, 9}. For each k, we ran our algorithm on each probabilistic network for 1200 seconds. SCIP was also run for 1200 seconds on each network. ... We used restart-based local search to perform search over the value assignments to S. ... We implemented the greedy MCKP algorithm of [14] to solve Ps. ... We used the max-degree heuristic outlined in section 4.2 to select a minimal k-separator.