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..
Generalization Bounds for Model-based Algorithm Configuration
Authors: Zhiyang Chen, Hailong Yao, Xia Yin
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we present numerical results for our generalization bound using the empirical Rademacher complexity by Algorithm 2. Experiments are conducted on a 2.0GHz Intel CPU with 4 cores. AC settings. We perform experiments on the algorithm configuration of the SCIP integer programming (IP) solver. |
| Researcher Affiliation | Academia | 1Tsinghua University 2University of Science and Technology Beijing 3Key Laboratory of Advanced Materials and Devices for Post-Moore Chips, Ministry of Education of China |
| Pseudocode | Yes | Algorithm 1 The algorithmic framework of model-based algorithm configurators. |
| Open Source Code | No | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [No] Justification: We will release the codes upon the acceptance of the paper. |
| Open Datasets | No | We randomly synthesize instances of VLSI routing on a 20 20 grid graph with 5 to 10 nets. Each net randomly consists of 2 to 5 points. The length of each grid edge is set to be 1. Facility location is a classic operations research problem... We randomly synthesize instances of facility location with n [400, 500] customers and m [200, 300] facilities. |
| Dataset Splits | No | The parameters found by algorithm configurators are only evaluated on sampled problem instances from distribution D. Why do they perform well on other unseen instances from D? |
| Hardware Specification | Yes | Experiments are conducted on a 2.0GHz Intel CPU with 4 cores. |
| Software Dependencies | No | We perform experiments on the algorithm configuration of the SCIP integer programming (IP) solver. |
| Experiment Setup | Yes | We set the (hyper-)parameters of the configurator as follows: Tinit = 20, Titer = 10, mbag = 10, Q = 1000, Οmin = 0.1, Ο = 0.1, TMH = 105, and = 0.03. |