DTW-CNN
Aligns each IC curve to a pristine reference with dynamic time warping and feeds the alignment to a convolutional network that estimates degradation modes. Trained on 'alawa synthetic curves.
Model details
- Architecture
- DTW features + CNN
- Input
- IC curve (dQ/dV vs. voltage)
- Output
- LLI, LAM_PE, LAM_NE (%)
- Training data
- Synthetic
- Uncertainty
- Not reported
- Physics constraints
- None
Evaluation results
| Task | Split | Score |
|---|---|---|
| Degradation-mode diagnosis | In-distribution | RMSE 1.60 |
| Degradation-mode diagnosis | Cross-chemistry | RMSE 4.40 |
| Degradation-mode diagnosis | Noisy and partial curves | RMSE 2.20 |