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 Generating Monolithic and Model Reconciling Explanations in Probabilistic Scenarios
Authors: Stylianos Loukas Vasileiou, William Yeoh, Alessandro Previti, Tran Cao Son
JAIR 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experimental evaluations on various benchmarks demonstrate the effectiveness and scalability of our approach in generating explanations under uncertainty. ... We present algorithms for computing explanations in Section 6, and experimentally evaluate them on a set of benchmarks in Section 7. |
| Researcher Affiliation | Collaboration | STYLIANOS LOUKAS VASILEIOU , New Mexico State University, United States WILLIAM YEOH, Washington University in St. Louis, United States ALESSANDRO PREVITI, Ericsson Research, Sweden TRAN CAO SON, New Mexico State University, United States |
| Pseudocode | Yes | Algorithm 1: monolithic-explanation(KB,𝜑) Algorithm 2: model-reconciling-explanation(KB𝛼, KBℎ 𝛼,𝜑) Algorithm 3: probabilistic-monolithic-explanation(B,𝜑, ˆ𝑘) Algorithm 4: probabilistic-model-reconciling-explanation(KB𝛼, Bℎ,𝜑, ˆ𝑘) |
| Open Source Code | Yes | Code repository: https://github.com/YODA-Lab/Probabilistic-Monolithic-Model-Reconciling-Explanations. |
| Open Datasets | Yes | Classical Planning Problems: We encoded classical planning problems from the International Planning Competition (IPC) in the style of Kautz et al. (1996), and used them as knowledge bases. ... Random CNF Problems: We generated random CNF formulae as knowledge bases using CNFgen (Lauria et al. 2017). |
| Dataset Splits | No | The paper mentions using problem instances from the International Planning Competition (IPC) and generating random CNF formulae using CNFgen, but it does not specify any training/test/validation splits for these problem instances or generated data. |
| Hardware Specification | Yes | Experiments were conducted on a system equipped with an M1 Max processor and 32GB of memory. |
| Software Dependencies | No | The algorithms were implemented in Python, utilizing the Py SAT toolkit (Ignatiev, Morgado, et al. 2018) for SAT solving, MCS/MUS finding, weighted Max SAT, and minimal hitting set computations. The specific versions for Python or Py SAT toolkit are not mentioned. |
| Experiment Setup | Yes | The time limit for all experiments was set to 500𝑠. |