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].
On Quantifying Literals in Boolean Logic and its Applications to Explainable AI
Authors: Adnan Darwiche, Pierre Marquis
JAIR 2021 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Theoretical | We formalized and studied the universal quantification of literals in Boolean logic, together with its applications to explainable AI. Our treatment was based on the novel notion of boundary models, which stands to have implications on the study of Boolean logic beyond quantification. A major contribution of our work is the interpretation of universal quantification as a selection process, which we hope will expand the applications of this form of quantification in AI and beyond. Another major contribution is the complexity results relating to the computation of existential and universal quantification on various logical forms. As to explainable AI, we have shown how to analyze classifiers and their decisions through the systematic construction of Boolean formulas, using universal literal and variable quantification. |
| Researcher Affiliation | Academia | Adnan Darwiche EMAIL Computer Science Department, UCLA, Los Angeles, CA 90095 USA Pierre Marquis EMAIL CRIL, Universit e d Artois & CNRS, Institut Universitaire de France F-62307, Lens Cedex, France |
| Pseudocode | No | The paper describes procedures and rules in prose and formal logic (e.g., Proposition 21, 22, 23 describe algorithms or methods), but it does not contain any explicitly labeled 'Pseudocode' or 'Algorithm' blocks or structured code-like procedures. |
| Open Source Code | No | The paper does not contain any statements about making code available, nor does it provide links to source code repositories for the methodology described. |
| Open Datasets | No | The paper uses conceptual examples of 'loan classifier' and 'admission classifier' but does not mention the use of any specific publicly available datasets for empirical evaluation. Footnote 1 mentions 'machine learning classifiers with discrete features using Boolean formulas...learned from data' but does not specify any dataset used in the paper's own analysis. |
| Dataset Splits | No | The paper does not mention any datasets or experimental evaluation, therefore, there is no information about training/test/validation dataset splits. |
| Hardware Specification | No | The paper is theoretical and does not describe any experiments that would require specific hardware. Therefore, no hardware specifications are provided. |
| Software Dependencies | No | The paper is theoretical and does not describe any implemented systems or experiments that would require specific software dependencies with version numbers. |
| Experiment Setup | No | The paper is theoretical and focuses on formalisms and logical properties; it does not describe an experimental setup with hyperparameters or training configurations. |