Learnable Weight Graph Neural Network for River Ice Classification
Bibliographic record
Abstract
Monitoring river ice is crucial for planning safe navigation routes, with ice–water classification being one of the most important tasks in ice mapping. While high-resolutions satellite imagery, such as synthetic aperture radar (SAR), is well-suited to this task, manual interpretation of these data is challenging due to the large data volume. Machine learning approaches are suitable methods to overcome this; however, training the models might not be time-effective when the desired result is a narrow structure, such as a river, within a large image. To address this issue, we proposed a model incorporating a graph neural network (GNN), called learnable weights graph convolution network (LWGCN). Focusing on the winters of 2017–2021 with emphasis on the Beauharnois Canal and Lake St Lawrence regions of the Saint Lawrence River. The model first converts the SAR image into graph-structured data using simple linear iterative clustering (SLIC) to segment the SAR image, then connecting the centers of each superpixel to form graph-structured data. For the training model, the LWGCN learns the weights on each edge to determine the relationship between ice and water. By using the graph-structured data as input, the proposed model training time is eight times faster, compared to a convolution neural network (CNN) model. Our findings also indicate that the LWGCN model can significantly enhance the accuracy of ice and water classification in SAR imagery.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".