Counting Linear Extensions in Practice: MCMC Versus Exponential Monte Carlo

Authors: Topi Talvitie, Kustaa Kangas, Teppo Niinimäki, Mikko Koivisto

AAAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental This work presents an empirical evaluation of the state-of-the-art schemes and investigates a number of ideas to enhance their performance. ... Through the empirical evaluation, we show that approximate counting is feasible on posets up to a few hundred elements.
Researcher Affiliation Academia Topi Talvitie Dept. of Computer Science University of Helsinki topi.talvitie@helsinki.fi Kustaa Kangas Dept. of Computer Science Aalto University juho-kustaa.kangas@aalto.fi Teppo Niinim aki Dept. of Computer Science Aalto University teppo.niinimaki@aalto.fi Mikko Koivisto Dept. of Computer Science University of Helsinki mikko.koivisto@helsinki.fi
Pseudocode No The paper describes algorithms verbally and with illustrative diagrams (e.g., Figure 2) but does not include formal pseudocode blocks or algorithms.
Open Source Code Yes We will make these instances as well as all algorithm implementations publicly available.1 (footnote refers to github.com/ttalvitie/le-counting-practice)
Open Datasets Yes Posets extracted from benchmark Bayesian networks, obtained from the Bayesian Network Repository (www.cs. huji.ac.il/ galel/Repository). We generated five posets from each instance class and size between 8 and 512 elements. We will make these instances as well as all algorithm implementations publicly available.1 (footnote refers to github.com/ttalvitie/le-counting-practice)
Dataset Splits No The paper does not mention specific training/validation/test dataset splits. It evaluates algorithms on generated instances and benchmark posets without describing such partitioning for model training.
Hardware Specification No Each run was limited to 24 hours of CPU time and 8 GB of RAM; all running times were measured. No specific CPU models, GPU models, or detailed hardware specifications are provided.
Software Dependencies No No specific software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or solvers with version numbers) are mentioned in the paper.
Experiment Setup Yes Each algorithm was instantiated to produce a (1, 1/4)-approximation for the number of linear extensions, i.e., an estimate within a factor of 2 with probability at least 3/4. In the experiments we set α = 0.1, β = 10, and κ = 5, observed to perform well on average.