Automated damage-sensitive feature extraction using unsupervised convolutional neural networks
Bibliographic record
Abstract
Many convolutional neural networks (CNN) –based approaches were proposed and applied to detect damage in various civil structures in recent years. Usually, the training process of the classical CNN requires a large number of labeled data which is from the monitored structure in undamaged and various damaged scenarios. However, it is impractical to acquire sufficient data that can be exactly labeled with damaged from the infrastructures in service as training data. Thus, we propose a novel unsupervised CNN-based approach to automatically extract optimal feature representations from the unlabeled data in a single class. In the case study, a known dataset from an undamaged scenario is used to train CNN and a dataset from an unknown scenario is used to test the trained CNN. The proposed approach in unsupervised learning is capable of extracting feature representations from the raw acceleration signals that are sensitive to the presence of damage. Then, the extracted damage-sensitive features are fed into a one-class support vector machine (OC-SVM) for novelty detection. The feature set from the undamaged dataset is taken as training dataset to train the OC-SVM, and the extracted features from the unknown dataset are used for testing. In order to verify the effectiveness of the proposed approach in structural damage localization, a number of accelerometers are used to acquire sufficient raw acceleration data from a lab-scale steel bridge, and the preliminary experimental results show that the proposed novel CNN-based approach performs very well in damage localization.
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) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| 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".