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..
S-SOS: Stochastic Sum-Of-Squares for Parametric Polynomial Optimization
Authors: Licheng Zhu, Mathias Oster, Yuehaw Khoo
NeurIPS 2024 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present two numerical studies of S-SOS demonstrating its use in applications. The first study (Section 3.1) numerically tests how the optimal values of the SDP Equation (2) p 2s converge to p of the original primal Equation (1) as we increase the degree. The second study (Section 3.2) evaluates the performance of S-SOS for solution extraction and uncertainty quantification in various sensor network localization problems. |
| Researcher Affiliation | Academia | Richard L. Zhu Department of Computational and Applied Mathematics University of Chicago Chicago, IL 60637; Mathias Oster Institute of Geometry and Practical Mathematics RWTH Aachen Aachen, Germany; Yuehaw Khoo Department of Statistics University of Chicago Chicago, IL 60637 |
| Pseudocode | Yes | Algorithm 1 Monte Carlo Point Optimization (MCPO) |
| Open Source Code | Yes | Our code has been provided with our submission and a cleaned-up version will be released publicly once the preprint is public. |
| Open Datasets | No | Once these are specified, we generate a random problem instance by sampling X Uniform( 1, 1)n, A Uniform( 1, 1)d. The paper generates its own data and does not provide public access information for it. |
| Dataset Splits | No | The paper describes generating random problem instances and evaluating performance, but does not specify any train/validation/test dataset splits or their sizes. |
| Hardware Specification | Yes | The SDPs are formulated with the help of Sym Py [34] and solved using CVXPY [35, 36] and Mosek [37] on a server with two Intel Xeon 6130 Gold processors (32 physical cores total) and 256GB of RAM. |
| Software Dependencies | No | The SDPs are formulated with the help of Sym Py [34] and solved using CVXPY [35, 36] and Mosek [37]... Specific version numbers for SymPy and CVXPY are not explicitly provided in the experimental setup description. |
| Experiment Setup | Yes | A given SNL problem type is specified by a spatial dimension ℓ, N sensors, K anchors, a sensing radius r (0, 2ℓ), a noise type (linear), and anchor type (soft penalty or hard equality). |