Approximate inference of marginals using the IBIA framework
Authors: Shivani Bathla, Vinita Vasudevan
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Results for several benchmark sets from recent UAI competitions show that our method gives either better or comparable accuracy than existing variational and sampling based methods, with smaller runtimes. |
| Researcher Affiliation | Academia | Shivani Bathla Department of Electrical Engineering Indian Institute of Technology Madras India, 600036 ee13s064@ee.iitm.ac.in Vinita Vasudevan Department of Electrical Engineering Indian Institute of Technology Madras India, 600036 vinita@ee.iitm.ac.in |
| Pseudocode | No | The paper describes the algorithm steps in paragraph form and through diagrams, but it does not present a formal pseudocode or algorithm block. |
| Open Source Code | No | The paper states that IBIA has been implemented in Python3 but does not provide any information about its public availability or a link to its source code. |
| Open Datasets | Yes | Benchmarks: We used the benchmark sets included in UAI repository [Ihler, 2006] and the Bayesian network repository [Scutari, 2007]. |
| Dataset Splits | No | The paper references standard benchmark datasets but does not explicitly describe specific train/validation/test dataset splits, percentages, or methodologies for partitioning the data. |
| Hardware Specification | Yes | All experiments were carried out on an Intel i9-12900 Linux system running Ubuntu 22.04. |
| Software Dependencies | No | The paper mentions software like Lib DAI and Merlin and programming languages like Python3 and C++, but it does not specify version numbers for these software components or libraries. |
| Experiment Setup | Yes | For IBIA, we set the maximum clique size bound mcsp to 20 (referred to as IBIA20 ) when the time limit is 2 min and we set it to 23 (referred to as IBIA23 ) when the time limit is 20 min. mcsim is empirically chosen as 5 less than mcsp. |