Research map / Time series / Causal and generative time series
NeurOCNN: A Neural-Operator-Based Model for Physiological Time Series
Daya Kumar, Uday Devulapalli, Aarat Satsangi, Apurva Narayan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace neural-operator backbones (FNO, U-NO, SNO, LocalNO) with spline convolution backbone | Fourier projection pooling model · ISRUC | accuracy (%): 70.99 -> 75.22 | Improved |
| replace neural-operator backbones with spline convolution backbone | Fourier projection pooling model · HMC | accuracy (%): 69.32 -> 72.88 | Improved |
| use NeurOCNN instead of ResNet | ECG classification · ECG | accuracy (%): 75.75 -> 76.08 | Improved |
| use NeurOCNN instead of U-Sleep | sleep staging · ISRUC | accuracy (%): 75.68 -> 75.22 | No clear effect |
| use shorter Fourier pooling segments (5 s, M=100) vs 30 s (M=600) at matched fmax (ablation) | NeurOCNN · ISRUC | accuracy (%): 75.89 -> 76.03 | No clear effect |
| use shorter Fourier pooling segments (ablation) | NeurOCNN · ISRUC | epoch time (s): 185.2 -> 153.5 | Improved |
| segment duration 15 s, M=300 vs 5 s, M=100 (ablation) | NeurOCNN · ISRUC | accuracy (%): 76.03 -> 74.46 | Hurt |
| increase spline control points on very short kernels (0.1 s) (ablation) | SplineConv1d · Sleep-EDF (EEG) | accuracy: 0.566 -> 0.264 (5 -> 25 points, per Figure 4 column) | Hurt |
| use moderate kernel duration (0.25-0.75 s) and 10-20 control points (ablation) | SplineConv1d · Sleep-EDF (EEG) | accuracy | Improved |
| increase number of control points at fixed kernel duration (ablation) | SplineConv1d · Sleep-EDF (EEG) | worst-case accuracy drop under sampling-rate shift | Hurt |
| add spline-kernel continuous convolution with Fourier pooling for noise robustness | NeurOCNN · ISRUC (noise at 5 dB SNR) | accuracy change | Improved |
Closest papers by meaning
- Neuroformer: Multimodal and Multitask Generative Pretraining for Brain Data (ICLR 2024 · similarity 0.85)
- POCO: Scalable Neural Forecasting through Population Conditioning (NeurIPS 2025 · similarity 0.85)
- BrainLM: A foundation model for brain activity recordings (ICLR 2024 · similarity 0.84)
- Learning semilinear neural operators: A unified recursive framework for prediction and data assimilation. (ICLR 2024 · similarity 0.84)
- NeurIPT: Foundation Model for Neural Interfaces (NeurIPS 2025 · similarity 0.84)
- NOBLE - Neural Operator with Biologically-informed Latent Embeddings to Capture Experimental Variability in Biological Neuron Models (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
