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..

Effective Neural Approximations for Geometric Optimization Problems

Authors: Samantha Chen, Oren Ciolli, Anastasios Sidiropoulos, Yusu Wang

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical results on diverse point-cloud datasets demonstrate the practical performance of our models. Section 5 is titled "Experiments", and numerous figures and tables throughout the paper (e.g., Figures 2, 6, and Tables 1, 2) present quantitative results, comparisons, and performance metrics.
Researcher Affiliation Academia Samantha Chen Computer Science and Engineering University of California, San Diego EMAIL Oren Ciolli Computer Science and Engineering University of California, San Diego EMAIL Anastasios Sidiropolous Computer Science and Engineering University of Illinois, Chicago EMAIL Yusu Wang Halıcıo glu Data Science Institute University of California, San Diego EMAIL
Pseudocode Yes Algorithm 1 Relaxed-ε-kernels; Algorithm 2 Computation of ϵ-kernels
Open Source Code Yes Our code is publically available1. 1https://github.com/chens5/coreset-nn
Open Datasets Yes Our experiments use (i) two synthetic datasets Gaussian Mixture... and (ii) two real datasets, SQUID [26] ... and the 3D Model Net [43]. ... The real datasets are freely available online.
Dataset Splits Yes Table 3: Dataset details for approximating relaxed-ε-kernels. |Ptrain| and |Ptest| refer to the size of the train and test point clouds, respectively. ... # train point clouds ... # test point clouds. Table 4: Dataset details for extent measure approximation tasks. |Ptrain| and |Ptest| refer to the size of the train and test point clouds, respectively. ... # train point clouds ... # test point clouds.
Hardware Specification Yes All models are implemented in Py Torch and trained on 8 NVIDIA RTX A6000 GPUs.
Software Dependencies No All models are implemented in Py Torch and trained on 8 NVIDIA RTX A6000 GPUs. ... Additionally, all models are trained using the ADAM optimizer [20] provided in Py Torch ... for MEA, we use the quadratic programming formulation implemented in the Computational Geometry Algorithms Library (CGAL) [35]. The paper mentions software like PyTorch, ADAM optimizer, and CGAL, but does not specify their version numbers.
Experiment Setup Yes All models are implemented in Py Torch and trained on 8 NVIDIA RTX A6000 GPUs. Our code is publically available1. Additionally, all models are trained using the ADAM optimizer [20] provided in Py Torch and trained using a learning rate of 0.001. ... Relaxed-ε-kernel networks are trained for 200 epochs while all extent-measure models are trained for 500 epochs. ... The depth of both the encoder and decoder MLPs are sampled from {2, 3, 4} and the width from {64, 128, 256}.