Beyond Parity Constraints: Fourier Analysis of Hash Functions for Inference

Authors: Tudor Achim, Ashish Sabharwal, Stefano Ermon

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

Reproducibility Variable Result LLM Response
Research Type Experimental Using this framework we introduce two new classes of hash functions for probabilistic inference and model counting that show promising performance on synthetic and real-world benchmarks.
Researcher Affiliation Collaboration Tudor Achim TACHIM@CS.STANFORD.EDU Stanford University, 353 Serra Mall, Stanford, CA 94305 Ashish Sabharwal ASHISHS@ALLENAI.ORG Allen Institute for Artificial Intelligence, 2157 N Northlake Way, Seattle, WA 98103 Stefano Ermon ERMON@CS.STANFORD.EDU Stanford University, 353 Serra Mall, Stanford, CA 94305
Pseudocode Yes Algorithm 1 THF-WISH-LB(w(x), n, T, Hβ)
Open Source Code No The paper does not provide an explicit statement or link to its own open-source code for the described methodology.
Open Datasets No The paper mentions "grid Ising models" and "real-world CNF formulas that encode problems in a wide range of domains (latin squares, Langford s problem, logistic planning, and hardware verification)". While these are known problem types, the paper does not provide concrete access information (e.g., URLs, DOIs, or specific citations to dataset repositories) for the exact instances used.
Dataset Splits No The paper does not specify training, validation, or test dataset splits.
Hardware Specification No The paper mentions execution times (e.g., "terminated within three seconds") and discusses a "solver" but does not specify any particular hardware components (e.g., CPU, GPU models, or memory) used for the experiments.
Software Dependencies No The paper mentions "lib DAI inference algorithm library (Mooij, 2010)" and "modern SAT solvers", but it does not provide specific version numbers for these software components.
Experiment Setup Yes We limit each MAP query in the inner loop of THF-WISH-LB and SPARSEWISH to five minutes. Because the theoretically-motivated density of variables in the sparse parity constraints used by SPARSE-WISH was still too high for the solver to find informative solutions within five minutes, we relaxed the constraint density to 5% of the variables for all runs.