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..
Siegel Neural Networks
Authors: Xuan Son Nguyen, Aymeric Histace, Nistor Grozavu
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | This section reports results of our experiments on the radar clutter classification and node classification tasks. For further details, please refer to Appendix 1 in which we present more experimental results on human action recognition and Riemannian generative modeling. |
| Researcher Affiliation | Academia | Xuan Son Nguyen Aymeric Histace Nistor Grozavu ETIS, UMR 8051, CY Cergy Paris University, ENSEA, CNRS EMAIL EMAIL |
| Pseudocode | No | The paper describes algorithms and methods in text and mathematical formulations but does not contain a clearly labeled 'Pseudocode' or 'Algorithm' block. |
| 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: The datasets used for our experiments will be made available upon acceptance of the paper. In the main paper and Appendix, we already give details on our experimental settings and implementation which would be sufficient to reproduce our experimental results. |
| Open Datasets | Yes | Radar clutter classification aims at recognizing different types of radar clutter which is the information recorded by a radar related to seas, forests, fields, cities and other environmental elements surrounding the radar [10]. Due to the scarcity of publicly available radar datasets for the task, our experiments are performed using simulated radar signals1 which are commonly assumed to be stationary centered autoregressive (AR) Gaussian time series [3, 4, 6, 10]. The AR model is given by ... 1https://github.com/nguyenxuanson10/synthetic-data ... We perform node classification experiments on Glass, Iris, and Zoo datasets from the UCI Machine Learning Repository [14]2. Like [27], our main aim is to demonstrate the applicability of our approach on Siegel spaces, and we do not necessarily seek state-of-the-art results for the target task. ... 2https://archive.ics.uci.edu/datasets |
| Dataset Splits | No | The paper mentions running experiments |
| Hardware Specification | No | The NeurIPS Paper Checklist states that hardware information is provided in Appendix 1, but Appendix 1 is not included in the provided text. The main body of the paper does not specify any hardware details such as GPU/CPU models or specific compute resources. |
| Software Dependencies | No | The paper does not explicitly mention any software dependencies with specific version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | Each of our networks consists of an FC (AFC or DFC) layer and a MLR layer built on the distance in Theorem 3.8. The sizes of the parameter b in the DFC layer are set to 3 2, 4 3, 5 3, and 6 4 for the experiments on datasets 1, 2, 3, and 4, respectively. We compare our approach to the following methods: ... We use default settings for SPD models as in the original papers (see Appendix 1.1). ... In our experiments, the embedding dimension is set to 6. |