Generalization Of Model Research Articles

Multichannel Adaptive Data Mixture Augmentation for Graph Neural Networks

Graph neural networks (GNNs) have demonstrated significant potential in analyzing complex graph-structured data. However, conventional GNNs encounter challenges in effectively incorporating global and local features. Therefore, this paper introduces a novel approach for GNN called multichannel adaptive data mixture augmentation (MAME-GNN). It enhances a GNN by adopting a multi-channel architecture and interactive learning to effectively capture and coordinate the interrelationships between local and global graph structures. Additionally, this paper introduces the polynomial–Gaussian mixture graph interpolation method to address the problem of single and sparse graph data, which generates diverse and nonlinear transformed samples, improving the model's generalization ability. The proposed MAME-GNN is validated through extensive experiments on publicly available datasets, showcasing its effectiveness. Compared to existing GNN models, the MAME-GNN exhibits superior performance, significantly enhancing the model's robustness and generalization ability.

Research into the Applications of a Multi-Scale Feature Fusion Model in the Recognition of Abnormal Human Behavior.

Due to the increasing severity of aging populations in modern society, the accurate and timely identification of, and responses to, sudden abnormal behaviors of the elderly have become an urgent and important issue. In the current research on computer vision-based abnormal behavior recognition, most algorithms have shown poor generalization and recognition abilities in practical applications, as well as issues with recognizing single actions. To address these problems, an MSCS-DenseNet-LSTM model based on a multi-scale attention mechanism is proposed. This model integrates the MSCS (Multi-Scale Convolutional Structure) module into the initial convolutional layer of the DenseNet model to form a multi-scale convolution structure. It introduces the improved Inception X module into the Dense Block to form an Inception Dense structure, and gradually performs feature fusion through each Dense Block module. The CBAM attention mechanism module is added to the dual-layer LSTM to enhance the model's generalization ability while ensuring the accurate recognition of abnormal actions. Furthermore, to address the issue of single-action abnormal behavior datasets, the RGB image dataset RIDS (RGB image dataset) and the contour image dataset CIDS (contour image dataset) containing various abnormal behaviors were constructed. The experimental results validate that the proposed MSCS-DenseNet-LSTM model achieved an accuracy, sensitivity, and specificity of 98.80%, 98.75%, and 98.82% on the two datasets, and 98.30%, 98.28%, and 98.38%, respectively.

Drug–target interaction prediction through fine-grained selection and bidirectional random walk methodology

The study of drug–target interaction plays an important role in the process of drug development. The subject of DTI forecasting has advanced significantly in the last several years, yielding numerous significant research findings and methodologies. Heterogeneous data sources provide richer information and comprehensive perspectives for drug–target interaction prediction, so many existing methods rely on heterogeneous networks, and graph embedding technology becomes an important technology to extract information from heterogeneous networks. These approaches, however, are less concerned with potential noisy information in heterogeneous networks and more focused on the extent of information extraction in those networks. Based on this, a potential DTI predictive network model called FBRWPC is proposed in this paper. It uses a fine-grained similarity selection program to first integrate similarity on similar networks and then a bidirectional random walk graph embedding learning method with restart to obtain an updated drug target interaction matrix. Through the use of similarity selection and fine-grained selection similarity integration, the framework can effectively filter out the noise present in heterogeneous networks and enhance the model's prediction performance. The experimental findings demonstrate that, even after being split up into four distinct types of data sets, FBRWPC can still retain great prediction performance, a sign of the model's resilience and good generalization.

A Hierarchical Multi-Task Learning Framework for Semantic Annotation in Tabular Data.

To optimize the utilization and analysis of tables, it is essential to recognize and understand their semantics comprehensively. This requirement is especially critical given that many tables lack explicit annotations, necessitating the identification of column types and inter-column relationships. Such identification can significantly augment data quality, streamline data integration, and support data analysis and mining. Current table annotation models often address each subtask independently, which may result in the neglect of constraints and contextual information, causing relational ambiguities and inference errors. To address this issue, we propose a unified multi-task learning framework capable of concurrently handling multiple tasks within a single model, including column named entity recognition, column type identification, and inter-column relationship detection. By integrating these tasks, the framework exploits their interrelations, facilitating the exchange of shallow features and the sharing of representations. Their cooperation enables each task to leverage insights from the others, thereby improving the performance of individual subtasks and enhancing the model's overall generalization capabilities. Notably, our model is designed to employ only the internal information of tabular data, avoiding reliance on external context or knowledge graphs. This design ensures robust performance even with limited input information. Extensive experiments demonstrate the superior performance of our model across various tasks, validating the effectiveness of unified multi-task learning framework in the recognition and comprehension of table semantics.

Enhanced multi-constraint dung beetle optimization-kernel extreme learning machine for lithium-ion battery state of health estimation with adaptive enhancement ability

Accurately estimating the state of health (SOH) of lithium batteries is a critical and challenging task in battery management systems. Data-driven models are widely used for SOH estimation but still suffer from the difficulty of balancing speed, accuracy, and adaptability. Therefore, this study constructs the dung beetle optimization algorithm to optimize the kernel extreme learning machine model. This paper addresses the issues of long iteration time and mismatches in kernel function mapping in data-driven models. To improve the model's generality, an adaptive learning kernel function is designed to complement the polynomial kernel function and form a joint function. This joint function is then introduced into a single implicit-layer extreme learning machine, which achieves fast speed and strong adaptive capability. To enhance the algorithmic parameter search capability, the optimal Latin hypercube idea, and the Osprey algorithm's global exploration strategy are introduced, which effectively improves the algorithm's global search capability. Additionally, it successfully regulated the positional update through the design of the logarithmic weighting factor, which improved the local search and convergence capabilities of the algorithm. The experiment validates the effectiveness and rationality of the proposed model for advancing battery management system applications.

A comprehensive analysis of evaluating robustness and generalization ability of models in AES

A comprehensive analysis of evaluating robustness and generalization ability of models in AES

Construction of IRAK4 inhibitor activity prediction model based on machine learning.

Interleukin-1 receptor-associated kinase 4 (IRAK4) is a crucial serine/threonine protein kinase that belongs to the IRAK family and plays a pivotal role in Toll-like receptor (TLR) and Interleukin-1 receptor (IL-1R) signaling pathways. Due to IRAK4's significant role in immunity, inflammation, and malignancies, it has become an intriguing target for discovering and developing potent small-molecule inhibitors. Consequently, there is a pressing need for rapid and accurate prediction of IRAK4 inhibitor activity. Leveraging a comprehensive dataset encompassing activity data for 1628 IRAK4 inhibitors, we constructed a prediction model using the LightGBM algorithm and molecular fingerprints. This model achieved an R2 of 0.829, an MAE of 0.317, and an RMSE of 0.460 in independent testing. To further validate the model's generalization ability, we tested it on 90 IRAK4 inhibitors collected in 2023. Subsequently, we applied the model to predict the activity of 13,268 compounds with docking scores less than -9.503kcal/mol. These compounds were initially screened from a pool of 1.6 million molecules in the chemdiv database through high-throughput molecular docking. Among these, 259 compounds with predicted pIC50 values greater than or equal to 8.00 were identified. We then performed ADMET predictions on these selected compounds. Finally, through a rigorous screening process, we identified 34 compounds that adhere to the four complementary drug-likeness rules, making them promising candidates for further investigation. Additionally, molecular dynamics simulations confirmed the stable binding of the screened compounds to the IRAK4 protein. Overall, this work presents a machine learning model for accurate prediction of IRAK4 inhibitor activity and offers new insights for subsequent structure-guided design of novel IRAK4 inhibitors.

Correcting Optical Aberration via Depth-Aware Point Spread Functions.

Optical aberration is a ubiquitous degeneration in realistic lens-based imaging systems. Optical aberrations are caused by the differences in the optical path length when light travels through different regions of the camera lens with different incident angles. The blur and chromatic aberrations manifest significant discrepancies when the optical system changes. This work designs a transferable and effective image simulation system of simple lenses via multi-wavelength, depth-aware, spatially-variant four-dimensional point spread functions (4D-PSFs) estimation by changing a small amount of lens-dependent parameters. The image simulation system can alleviate the overhead of dataset collecting and exploiting the principle of computational imaging for effective optical aberration correction. With the guidance of domain knowledge about the image formation model provided by the 4D-PSFs, we establish a multi-scale optical aberration correction network for degraded image reconstruction, which consists of a scene depth estimation branch and an image restoration branch. Specifically, we propose to predict adaptive filters with the depth-aware PSFs and carry out dynamic convolutions, which facilitate the model's generalization in various scenes. We also employ convolution and self-attention mechanisms for global and local feature extraction and realize a spatially-variant restoration. The multi-scale feature extraction complements the features across different scales and provides fine details and contextual features. Extensive experiments demonstrate that our proposed algorithm performs favorably against state-of-the-art restoration methods.

DBAD: Dual branch reconstruction for industrial anomaly detection

AbstractReconstruction‐based methods are commonly used in industrial visual anomaly detection. They rely on a well‐reconstructed normal mode of the model. However, it is difficult to manage the boundary of generalization. The strength of the model's generalization capability can directly affect the fidelity of the reconstruction, resulting in the occurrence of false positives. To address the above challenges, a novel dual branch reconstruction anomaly detection approach is proposed to control the model generalization capability at two dimensions. It reconstructs abnormal images into normal ones by resolution recovery and denoising branches. Detection results are generated from their comparison. In addition, an innovative channel adjustment module is introduced to improve information exchange between branches. It uses multiple dilated convolutions for interactions over different scales. Simulation experiments demonstrate that the method outperforms most inspection methods on the MVTec AD and MVTec 3D‐AD datasets. It also shows good results on the self‐generated automotive paint scratches dataset of this study.

A Self-Supervised Few-Shot Semantic Segmentation Method Based on Multi-Task Learning and Dense Attention Computation.

Nowadays, autonomous driving technology has become widely prevalent. The intelligent vehicles have been equipped with various sensors (e.g., vision sensors, LiDAR, depth cameras etc.). Among them, the vision systems with tailored semantic segmentation and perception algorithms play critical roles in scene understanding. However, the traditional supervised semantic segmentation needs a large number of pixel-level manual annotations to complete model training. Although few-shot methods reduce the annotation work to some extent, they are still labor intensive. In this paper, a self-supervised few-shot semantic segmentation method based on Multi-task Learning and Dense Attention Computation (dubbed MLDAC) is proposed. The salient part of an image is split into two parts; one of them serves as the support mask for few-shot segmentation, while cross-entropy losses are calculated between the other part and the entire region with the predicted results separately as multi-task learning so as to improve the model's generalization ability. Swin Transformer is used as our backbone to extract feature maps at different scales. These feature maps are then input to multiple levels of dense attention computation blocks to enhance pixel-level correspondence. The final prediction results are obtained through inter-scale mixing and feature skip connection. The experimental results indicate that MLDAC obtains 55.1% and 26.8% one-shot mIoU self-supervised few-shot segmentation on the PASCAL-5i and COCO-20i datasets, respectively. In addition, it achieves 78.1% on the FSS-1000 few-shot dataset, proving its efficacy.

PEL-PVP: Application of plant vacuolar protein discriminator based on PEFT ESM-2 and bilayer LSTM in an unbalanced dataset

Plant vacuoles, play a crucial role in maintaining cellular stability, adapting to environmental changes, and responding to external pressures. The accurate identification of vacuolar proteins (PVPs) is crucial for understanding the biosynthetic mechanisms of intracellular vacuoles and the adaptive mechanisms of plants. In order to more accurately identify vacuole proteins, this study developed a new predictive model PEL-PVP based on ESM-2. Through this study, the feasibility and effectiveness of using advanced pre-training models and fine-tuning techniques for bioinformatics tasks were demonstrated, providing new methods and ideas for plant vacuolar protein research. In addition, previous datasets for vacuolar proteins were balanced, but imbalance is more closely related to the actual situation. Therefore, this study constructed an imbalanced dataset UB-PVP from the UniProt database，helping the model better adapt to the complexity and uncertainty in real environments, thereby improving the model's generalization ability and practicality. The experimental results show that compared with existing recognition techniques, achieving significant improvements in multiple indicators, with 6.08 %, 13.51 %, 11.9 %, and 5 % improvements in ACC, SP, MCC, and AUC, respectively. The accuracy reaches 94.59 %, significantly higher than the previous best model GraphIdn. This provides an efficient and precise tool for the study of plant vacuole proteins.

A deep learning framework for predicting endometrial cancer from cytopathologic images with different staining styles.

Endometrial cancer screening is crucial for clinical treatment. Currently, cytopathologists analyze cytopathology images is considered a popular screening method, but manual diagnosis is time-consuming and laborious. Deep learning can provide objective guidance efficiency. But endometrial cytopathology images often come from different medical centers with different staining styles. It decreases the generalization ability of deep learning models in cytopathology images analysis, leading to poor performance. This study presents a robust automated screening framework for endometrial cancer that can be applied to cytopathology images with different staining styles, and provide an objective diagnostic reference for cytopathologists, thus contributing to clinical treatment. We collected and built the XJTU-EC dataset, the first cytopathology dataset that includes segmentation and classification labels. And we propose an efficient two-stage framework for adapting different staining style images, and screening endometrial cancer at the cellular level. Specifically, in the first stage, a novel CM-UNet is utilized to segment cell clumps, with a channel attention (CA) module and a multi-level semantic supervision (MSS) module. It can ignore staining variance and focus on extracting semantic information for segmentation. In the second stage, we propose a robust and effective classification algorithm based on contrastive learning, ECRNet. By momentum-based updating and adding labeled memory banks, it can reduce most of the false negative results. On the XJTU-EC dataset, CM-UNet achieves an excellent segmentation performance, and ECRNet obtains an accuracy of 98.50%, a precision of 99.32% and a sensitivity of 97.67% on the test set, which outperforms other competitive classical models. Our method robustly predicts endometrial cancer on cytopathologic images with different staining styles, which will further advance research in endometrial cancer screening and provide early diagnosis for patients. The code will be available on GitHub.

Performance analysis of deep learning-based electric load forecasting model with particle swarm optimization

With the widespread application of deep learning technology in various fields, power load forecasting, as an important link in power system operation and planning, has also ushered in new opportunities and challenges. Traditional forecasting methods perform poorly when faced with the high uncertainty and complexity of power loads. In view of this, this paper proposes a power load forecasting model PSO-BiTC based on deep learning and particle swarm optimization. This model combines a temporal convolutional network (TCN) and a bidirectional long short-term memory network (BiLSTM), using TCN to process long sequence data and capture features and patterns in time series, while using BiLSTM to capture long-term and short-term dependencies. In addition, the particle swarm optimization algorithm (PSO) is used to optimize model parameters to improve the model's predictive performance and generalization ability. Experimental results show that the PSO-BiTC model performs well in power load forecasting. Compared with traditional methods, this model reduces the MAE (Mean Absolute Error) to 20.18, 17.57, 18.61 and 16.7 on four extensive data sets, respectively. It has been proven that it achieves the best performance in various indicators, with a low number of parameters and training time. This research is of great significance for improving the operating efficiency of the power system, optimizing resource allocation, and promoting carbon emission reduction goals in the urban building sector.

Bayesian optimization-based Model-Agnostic Meta-Learning: Application to predict maximum cyclic moment resistance of steel bolted T-stub connections

Accurately assessing the maximum moment resistance of steel bolted T-stub connections under cyclic loading is crucial for designing earthquake-resistant structures with such connections. Traditional methods based on design standards like ASCE 41-17 often lack precision. Recently, supervised machine learning techniques, particularly Artificial Neural Network (ANN), have been explored. However, conventional ANNs require substantial data for generalization, which is limited for steel bolted T-stub connections. To address these challenges, this study explores the feasibility of using the Model-Agnostic Meta-Learning (MAML) to predict the maximum cyclic moment resistance of steel bolted T-stub connections. MAML adapts task-specific model parameters rapidly and transfers knowledge across tasks to fine-tune global model parameters, potentially enhancing prediction accuracy with limited data. The MAML model is first optimized using Bayesian optimization with a Gaussian Process model to identify ideal hyperparameters. The optimized MAML model is then compared with two ANN models (one with optimized hyperparameters, another matching MAML's neural network architecture) and ASCE 41-17 method. Results demonstrate the optimized MAML model's superior generalization capabilities, offering a promising approach for steel bolted T-stub connections.

Development and validation of a deep learning-based framework for automated lung CT segmentation and acute respiratory distress syndrome prediction: a multicenter cohort study

Acute respiratory distress syndrome (ARDS) is a life-threatening condition with a high incidence and mortality rate in intensive care unit (ICU) admissions. Early identification of patients at high risk for developing ARDS is crucial for timely intervention and improved clinical outcomes. However, the complex pathophysiology of ARDS makes early prediction challenging. This study aimed to develop an artificial intelligence (AI) model for automated lung lesion segmentation and early prediction of ARDS to facilitate timely intervention in the intensive care unit. A total of 928 ICU patients with chest computed tomography (CT) scans were included from November 2018 to November 2021at three centers in China. Patients were divided into a retrospective cohort for model development and internal validation, and three independent cohorts for external validation. A deep learning-based framework using the UNet Transformer (UNETR) model was developed to perform the segmentation of lung lesions and early prediction of ARDS. We employed various data augmentation techniques using the Medical Open Network for AI (MONAI) framework, enhancing the training sample diversity and improving the model's generalization capabilities. The performance of the deep learning-based framework was compared with a Densenet-based image classification network and evaluated in external and prospective validation cohorts. The segmentation performance was assessed using the Dice coefficient (DC), and the prediction performance was assessed using area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. The contributions of different features to ARDS prediction were visualized using Shapley Explanation Plots. This study was registered with the China Clinical Trial Registration Centre (ChiCTR2200058700). The segmentation task using the deep learning framework achieved a DC of 0.734±0.137 in the validation set. For the prediction task, the deep learning-based framework achieved AUCs of 0.916 [0.858-0.961], 0.865 [0.774-0.945], 0.901 [0.835-0.955], and 0.876 [0.804-0.936] in the internal validation cohort, external validation cohort I, external validation cohort II, and prospective validation cohort, respectively. It outperformed the Densenet-based image classification network in terms of prediction accuracy. Moreover, the ARDS prediction model identified lung lesion features and clinical parameters such as C-reactive protein, albumin, bilirubin, platelet count, and age as significant contributors to ARDS prediction. The deep learning-based framework using the UNETR model demonstrated high accuracy and robustness in lung lesion segmentation and early ARDS prediction, and had good generalization ability and clinical applicability. This study was supported by grants from the Shanghai Renji Hospital Clinical Research Innovation and Cultivation Fund (RJPY-DZX-008) and Shanghai Science and Technology Development Funds (22YF1423300).

Assessment of Hull and Propeller Performance Degradation Based on TSO-GA-LSTM

Evaluating the degradation of hull and ship performance and exploring their degradation pathways is crucial for developing scientific and reasonable ship maintenance plans. This paper proposes a two-stage optimization (TSO) algorithm that combines the Genetic Algorithm (GA) and Long Short-Term Memory (LSTM) network, capable of simultaneously optimizing input features and model parameters to enhance the accuracy and generalization ability of speed prediction models. Additionally, a performance degradation assessment method based on speed loss is provided, aimed at evaluating the degradation of hull and propeller performance, as well as extracting the performance degradation paths. The results indicated that the proposed TSO-LSTM-GA algorithm significantly outperformed existing baseline models. Furthermore, the provided performance degradation assessment method demonstrated certain effectiveness on the target ship data, with a measured degradation rate of 0.00344 kn/d and a performance degradation of 9.569% over 478 days, corresponding to an annual speed loss of 1.257 kn.

A data-driven model for predicting fatigue performance of high-strength steel wires based on optimized XGBOOST

Fatigue damage to high-strength steel wires in cable-stayed bridges can significantly compromise bridge reliability. Previous studies have primarily focused on isolated factors such as corrosion rate or stress ratio, failing to capture the complex interactions among multiple variables. Consequently, a few researchers have developed data-driven models using machine learning methods. However, these unoptimized models exhibit limited prediction performance and convergence speed, and lack in-depth analysis of feature effects on the models. To address these shortcomings, this study combines the Gray Wolf Optimization (GWO) algorithm with the XGBoost model to create a data-driven approach for predicting the fatigue performance of high-strength steel wires under multiple influencing factors. The model is validated using both journal and experimental datasets. Results demonstrate that the proposed GWO-XGBoost model achieves high prediction accuracy (R2 = 0.94) and strong generalization ability, and exhibits rapid convergence in optimizing hyperparameters, outperforming BP neural networks and ridge regression models. Additionally, the study highlights the primary effects of wire mass loss, stress range, and average stress on fatigue life prediction, emphasizing the importance of preventing corrosion and managing stress levels to enhance wire fatigue performance. Analysis of model parameters reveals that alpha and colsample_bytree are critical for maintaining model stability and minimizing error rates, while maximum depth and learning rate significantly impact model complexity and convergence. This suggests that proper tuning of these parameters is essential for ensuring model robustness, efficiency, and generalization ability.

Development of ensemble learning techniques and sequential model-based optimization for enhancing the generalizability of shale wettability predictions

Quantifying the wettability of shales is important for reservoir exploration and evaluation, as well as CO2 storage. Conventional experimental measurements are time-consuming and costly, while novel machine learning (ML) algorithms are challenged to predict shale wettability with accuracy and generalizability. Therefore, this study introduces an advanced shale wettability prediction framework by fusing ensemble learning algorithm and automatic hyperparameters optimization schemes (AHOS). Specifically, we firstly collected data from various gas geo-storage conditions from the literature. Then, an ensemble learning algorithm was utilized to extract complex nonlinear relationships among influential variables and shale wettability. During the training process, multiple optimization schemes, including random search (RS), Bayesian optimization (BO), and sequential model-based optimization (SMBO), were used and compared to automatically find the optimal hyperparameters of a wettability prediction model. The results show that AHOS can automatically and efficiently obtain the optimal hyperparameters, thus improves model's generalization. In addition, the prediction model based on AHOS process can accurately predict shale wettability. Through sensitivity analysis, we also revealed that the TOC and gas type are the top two most important influencing factors, which provides insights for CO2 sequestration. Moreover, we validated that the CatBoost model outperforms other tree-based models in predicting shale wettability and the SMBO method can provide a more effective result. In summary, the proposed shale wettability prediction framework is promising and efficient, which contributes to the exploration, evaluation, and development of shale oil reservoirs, further benefit to CO2 storage.

Enhancing Image Classification Using Few-Shot Learning Prototypical Networks with ResNet-18: Detection, Accuracy Enhancement, and Optimization

Image classification, especially in scenarios with limited data, presents significant challenges. Few shot learning (FSL) aims to address these challenges by training models that can generalize from a few examples. This paper explores the integration of prototypical networks with ResNet-18 for feature extraction to enhance image classification accuracy. Prototypical networks are designed to create a prototype representation for each class, which can then be used to classify new examples based on their distance to these prototypes. By leveraging ResNet-18's powerful feature extraction capabilities, we aim to improve the quality of these prototypes, thereby enhancing classification performance.We propose various methods for accuracy enhancement and optimization, including hyperparameter tuning, regularization techniques, and advanced methods like attention mechanisms and metric learning. Hyperparameter tuning involves adjusting the model's parameters to find the optimal settings that yield the best performance. Regularization techniques, such as dropout and weight decay, help prevent overfitting and improve the model's generalization capabilities. Advanced methods like attention mechanisms can focus on the most relevant parts of the image, while metric learning aims to learn a distance metric that better reflects the similarities between images.Our experiments on datasets like Mini-ImageNet and Omniglot demonstrate significant improvements in classification performance. These datasets are commonly used benchmarks in the few-shot learning community, allowing us to compare our results with existing methods. The integration of prototypical networks with ResNet-18, along with the proposed optimization techniques, provides a robust approach for tackling the challenges of image classification in few-shot learning scenarios. Key Words: Few-shot learning, ResNet-18, Prototypical Networks.

Intelligent recognition of structural health state of EV lithium-ion Battery using transfer learning based on X-ray computed tomography

The lithium-ion battery (LIB) achieves wide applications on electric vehicles (EVs). The morphology of LIB electrode is highly correlated to the performance and safety of EV. Therefore, it is significant to monitor the structural health state of LIB to guarantee a reliable operation of EV. Currently, few research focused on the intelligent recognition of LIB structural health state. To bridge the gap, this paper proposes an image-based method to conduct structural health state recognition of LIB. The geometric deformation of LIB electrode is recorded by X-ray computed tomography (CT) images under different structure health state. Then a novel transfer learning network is incorporated to enhance the generalization ability of model. In the network, prior knowledge is first acquired from source domain data to initialize a convolutional autoencoder (CAE). Then the CAE is fine-tuned on CT images (target domain data), and a structural deviation index is established to reflect the structure health condition. Once the index exceeds the pre-set threshold, the alarm is triggered. The performance of the proposed method was evaluated on experimental data, and the results proved that the proposed method can perform structural health state monitoring effectively, and transfer learning can improve the generalization ability of detection model significantly.