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..
BioOSS: A Bio-Inspired Oscillatory State System with Spatio-Temporal Dynamics
Authors: Zhongju Yuan, Geraint Wiggins, Dick Botteldooren
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate Bio OSS on both synthetic and real-world tasks, demonstrating superior performance and enhanced interpretability compared to alternative architectures. We validate Bio OSS through extensive experiments on synthetic and real-world datasets, demonstrating both competitive performance and enhanced interpretability compared to baseline architectures. We evaluate the proposed Bio OSS model on two fundamental time series tasks: classification and prediction. |
| Researcher Affiliation | Academia | 1WAVES Research Group, Ghent University, Gent, Belgium 2AI Lab, Vrije Universiteit Brussel, Brussel, Belgium 3EECS, Queen Mary University of London, London, UK EMAIL, EMAIL, EMAIL |
| Pseudocode | Yes | Algorithm 1 illustrates the full architecture of the multi-layer Bio OSS model. While the main text focuses on a single-layer formulation, the full model stacks L such layers sequentially. |
| Open Source Code | Yes | The complete implementation code is provided in https://github. com/zjyuan1208/Bio OSS. |
| Open Datasets | Yes | This benchmark includes six multivariate time series datasets from the University of East Anglia (UEA) Multivariate Time Series Classification Archive (UEA-MTSCA) [Bagnall et al., 2018]: Eigen Worms (Worms), Self Regulation SCP1 (SCP1), Self Regulation SCP2 (SCP2), Ethanol Concentration (Ethanol), Heartbeat, and Motor Imagery (Motor). All datasets used in the experiments are publicly available, and appropriate citations are provided. |
| Dataset Splits | Yes | All models were trained under identical settings, with results averaged over five seeds and a 70:15:15 train/validation/test split. |
| Hardware Specification | Yes | Models were trained using the Adam optimizer and executed on a single NVIDIA Ge Force RTX 4090 GPU (24 GB memory). All experiments were conducted on the same GPU device to guarantee fair and direct comparability. |
| Software Dependencies | No | All experiments were conducted using JAX for classification tasks and Py Torch for prediction tasks to ensure consistency with established baselines, datasets, and evaluation protocols. |
| Experiment Setup | Yes | For each task, the learning rate was selected through a grid search to ensure optimal performance. A comprehensive summary of all hyperparameter configurations is provided in the Appendix Section D Table 4. Models were trained using the Adam optimizer and executed on a single NVIDIA Ge Force RTX 4090 GPU (24 GB memory). |