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 Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Towards Real-Time Approximate Counting
Authors: Yash Pote, Kuldeep S. Meel, Jiong Yang
AAAI 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In an evaluation over 2,247 instances, Approx MC7 solved 271 more instances and achieved a 2 speedup against Approx MC. [...] Section 4 provides an empirical evaluation of Approx MC7 against Approx MC, and finally, we conclude in Section 5. |
| Researcher Affiliation | Academia | 1 National University of Singapore 2 University of Toronto 3 Georgia Institute of Technology |
| Pseudocode | Yes | Algorithm 1: Approx MC(φ, ε, δ) [...] Algorithm 2: Approx MC7(φ, ε, δ) [...] Algorithm 3: Compute Iter(ε, δ) [...] Algorithm 4: Approx MC7Core(φ, ε) |
| Open Source Code | Yes | 1The resulting tool Approx MC7 is available open-source at https://github.com/meelgroup/approxmc |
| Open Datasets | Yes | We evaluated the runtime performance of Approx MC7 and Approx MC6 over a comprehensive set of 2,247 instances (Yang, Pote, and Meel 2024) [...] Yang, J.; Pote, Y.; and Meel, K. S. 2024. Benchmark used for AAAI25 paper: Towards Real-Time Approximate Counting. https://doi.org/10.5281/zenodo.14533501. |
| Dataset Splits | No | The paper uses a set of 2,247 instances for evaluation and discusses using exact model counter Ganak for comparison on a subset of 698 instances. However, it does not describe specific training/test/validation splits for machine learning models, as the work focuses on model counting problems rather than training predictive models. |
| Hardware Specification | Yes | We conducted our experiments on a high-performance compute cluster, with each node consisting of AMD EPYCMilan processor featuring 2 64 real cores and 512GB of RAM. |
| Software Dependencies | No | The paper mentions using 'an efficient pre-processor Arjun (Soos and Meel 2022)' but does not provide a specific version number for Arjun or any other software components used in the experiments. |
| Experiment Setup | Yes | In our experiments, we set δ = 0.2 and ε = 13 and used an efficient pre-processor Arjun (Soos and Meel 2022) to simplify the function. [...] We ran each job on a single core with a 100-second time limit and 4GB memory. |