Balanced residual distillation learning for 3D point cloud class-incremental semantic segmentation

Expert Systems with Applications, Год журнала: 2025, Номер unknown, С. 126399 - 126399

Опубликована: Янв. 1, 2025

Язык: Английский

---

Prior knowledge-informed multi-task dynamic learning for few-shot machinery fault diagnosis

Expert Systems with Applications, Год журнала: 2025, Номер unknown, С. 126439 - 126439

Опубликована: Янв. 1, 2025

Язык: Английский

---

Acoustic emission and electromagnetic radiation precursor signal identification and early warning of coal and gas outburst based on diffusion-semi-supervised classification method

Physics of Fluids, Год журнала: 2024, Номер 36(12)

Опубликована: Дек. 1, 2024

Gas outbursts in coal seams represent a severe and formidable hazard, posing significant threat to the safety of mining operations. The advanced early warning is crucial preventive measure against outbursts. Acoustic emission (AE) electromagnetic radiation (EMR) are monitoring techniques for gas However, during operations, interference signals from AE EMR may arise. Due impact these signals, use statistical indicators time-frequency feature analysis lead false alarms missed detections outburst warnings. advancement deep learning offers new methods intelligent identification risks. This article proposes an method detecting precursor conducting comprehensive index based on EMR. First, reconstruct signal using wavelet packet decomposition then process resulting with diffusion-semi-supervised classification algorithm, employing partially labeled train model risk By analyzing prominent EMR, establish Bayesian networks, thereby achieving findings suggest that question, which employs training dataset comprising 60% manually annotated data, proficient precisely identifying adept at range signals. It provides basis distinguished multi-level warning. research outcomes significantly enhance reliability offering effective seams, power manifestations, abnormal gas.

Язык: Английский

---

An interpretable deep feature aggregation framework for machinery incremental fault diagnosis

Advanced Engineering Informatics, Год журнала: 2025, Номер 65, С. 103189 - 103189

Опубликована: Фев. 15, 2025

Язык: Английский

---

FSPDF: Few-shot learning with progressive dual-domain feature fusion via self-supervised learning

Knowledge-Based Systems, Год журнала: 2025, Номер unknown, С. 113389 - 113389

Опубликована: Апрель 1, 2025

Язык: Английский

---

Semi-supervised Feature Contrast Incremental Learning Framework for Bearing Fault Diagnosis with Limited Labeled Samples

Applied Soft Computing, Год журнала: 2025, Номер unknown, С. 113172 - 113172

Опубликована: Апрель 1, 2025

Язык: Английский

---

Compact-sparse prototype calibration network for few-shot continual fault diagnosis of rotating machinery

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 155, С. 111099 - 111099

Опубликована: Май 12, 2025

Язык: Английский

---

Simplicial complexes graph convolution networks with higher-order features learning for limited samples diagnosis

Control Engineering Practice, Год журнала: 2025, Номер 162, С. 106391 - 106391

Опубликована: Май 13, 2025

Язык: Английский

---

Prototype space boundary alignment network for bearing continuous fault diagnosis under class-incremental scenarios

IEEE Sensors Journal, Год журнала: 2024, Номер 25(1), С. 1076 - 1085

Опубликована: Ноя. 14, 2024

Язык: Английский

---

Development of a Hierarchical Clustering Method for Anomaly Identification and Labelling of Marine Machinery Data

Journal of Marine Science and Engineering, Год журнала: 2024, Номер 12(10), С. 1792 - 1792

Опубликована: Окт. 9, 2024

The application of artificial intelligence models for the fault diagnosis marine machinery increased expeditiously within shipping industry. This relates to effectiveness in capturing patterns systems that are becoming more complex and where traditional methods is unfeasible. However, despite these advances, lack labelling data still a major concern due confidentiality issues, appropriate data, instance. In this study, method based on histogram similarity hierarchical clustering proposed as an attempt label distinct anomalies faults occur dataset so supervised learning can then be implemented. To validate methodology, case study main engine tanker vessel considered. results indicate preliminary option classify types may appear dataset, model achieved accuracy approximately 95% presented.

Язык: Английский

---