Transfer Learning from Wi-Fi Access Point Data to Air Quality Monitoring: A Spatial-Temporal Graph-Based Approach
Bibliographic record
Abstract
Transfer learning has emerged as a powerful framework for leveraging knowledge from a source task to improve efficiency in learning a target task, particularly when labeled data is limited. In this paper, we propose a novel framework that applies transfer learning from Wi-Fi Access Point (AP) prediction to air quality monitoring—two domains with underlying spatial-temporal dependencies. We employ a model combining Graph Neural Networks (GNNs) with temporal modules to capture these patterns, first training on Wi-Fi AP data and transferring its pre-trained weights to air quality forecasting. By fine-tuning the model on the target domain while partially freezing earlier layers, we retain learned spatial-temporal representations while reducing computational costs. Experimental results on real-world datasets demonstrate that our transfer learning approach achieves comparable accuracy to training from Baseline but with significantly reduced training time. Furthermore, the framework is applicable for transferring knowledge within the same do-main for instance, between distinct ensembles of air quality nodes-enabling efficient adaptation to new sensor deployments without full retraining. This highlights the practical utility of cross-domain and intra-domain transfer learning for resource-constrained spatio-temporal applications. Experimental results demonstrate that our transfer learning framework achieves accuracy comparable to training from Baseline while reducing training epochs by 51.7% (14 vs. 29 epochs) and computation time by 52.7% (17.51s vs. 37.05s), highlighting its potential for resource-constrained applications.
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 | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".