Multilayer Grad-CAM: An effective tool towards explainable deep neural networks for intelligent fault diagnosis

Li, Sinan; Li, Tianfu; Sun, Chuang; Yan, Ruqiang; Chen, Xuefeng

doi:10.1016/j.jmsy.2023.05.027

Cited by 34 publications

(6 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to the setup shown in Figure 6, various NN designs have been tried. In previous works, there have been successful applications of deep NN in classification or detection [19,[58][59][60][61]. Intriguingly, augmenting the quantity of fully connected layers in this work led to a decrement in the recognition rate, and various combinations failed to yield discernibly improved outcomes.…”

Section: Neural Networkmentioning

confidence: 71%

See 1 more Smart Citation

Classification of LED Packages for Quality Control by Discriminant Analysis, Neural Network and Decision Tree

Shim,

Kim

2024

Micromachines

View full text Add to dashboard Cite

This study investigates supervised learning to improve LED classification. A hardware system for testing was built. The data for learning were acquired and then analyzed to show their characteristics. An LED was tested, and the results were categorized into three defective LED groups and one normal LED group. Before classification, electrical and optical data were examined to identify their characteristics. To find out the best way for quality control, an ensemble of methods was used. First, the discriminant analysis using the validation data achieved a 77.9% true positive rate for normal products, inadequate for quality control. Second, neural network-based learning boosted this rate to 97.8%, but the 2.2% false negative rate remained problematic. Finally, a binary decision tree was constructed, achieving a 99.4% true positive rate from just 14 splits, proving highly effective in product classification. The training time was measured as 8.1, 18.2 and 8.2 s for discriminant analysis, neural network and decision tree, respectively. This work has found the binary decision tree is advantageous considering both learning and classification efficiencies.

show abstract

Section: Neural Networkmentioning

confidence: 71%

Section: Neural Networkmentioning

confidence: 99%

Classification of LED Packages for Quality Control by Discriminant Analysis, Neural Network and Decision Tree

Shim,

Kim

2024

Micromachines

View full text Add to dashboard Cite

show abstract

“…A smaller MMD value means small feature distribution difference. Generally, MMD can be expressed as follows [45]: To further interpret the fault diagnosis procedure of MJA-Net, Layer CAM [46,47] are utilized to visualize the segments of the original input vibration signal and acoustic signal. When a given signal S is input to the classifier f, the predicted score y c of the target category c can be obtained by…”

Section: Features Analysismentioning

confidence: 99%

A multi-modal joint attention network for vibro-acoustic fusion diagnosis of engines

Zhu,

Zhang,

Wang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Deep learning has proven to be effective in diagnosing faults in power machinery and its diagnosis performance relies on a sufficient data set. In practice, a well-labeled data set with sufficient samples is very rare, especially for those machinery running in varying loading cases. The situation is particularly pronounced for multi-cylinder internal combustion engines, where the excitations from cylinders interact with significant background noise, and different data distributions are complicated. To tackle these issues, we propose a novelty multi-modal joint attention network (MJA-Net) for fusing the vibration and acoustic signals for diagnosing multiple faults. In MJA-Net, feature maps from both modalities are input separately into the convolutional module to learn independent features, and joint attention module (JAM) is utilized to enhance the vibro-acoustic information interaction and distribution consistency across modalities. The analysis of multiple loads vibro-acoustic experimental data shows that MJA-Net has a superior classification performance in limited sample tasks, compared to the single-modal methods. Furthermore, MJA-Net outperforms other fusion methods on average accuracy at 97.65%, as well as feature representativeness, and vibro-acoustic feature consistency across loads. JAM has superior diagnosis performance to other alternative modules. The class activation maps (CAM) generated by the Layer CAM highlight the key impact components related to the engine working mechanisms, providing valuable insight into MJA-Net’s interpretation for multi-fault recognition.

show abstract

“…Preprocessing methods, also termed modelbased methods, typically employ interpretable models like AE [17], CAE [18] and VAE [19] as a foundation for CNNs. Post hoc interpretability, such as Grad-CAM [20], Multilayer Grad-CAM [21], and time-frequency maps [22] extract crucial information from original signals after the model is built. Mey and Neufeld [23] visualize weights and orders.…”

Section: Introductionmentioning

confidence: 99%

WBUN: an interpretable convolutional neural network with wavelet basis unit embedded for fault diagnosis

Gao,

Zhang,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Convolutional Neural Network (CNN) is extensively applied in mechanical system fault diagnosis. However, the absence of transparent decision mechanisms in CNNs hinders credibility. To address these challenges, this paper proposes an interpretable wavelet basis unit convolutional network (WBUN). This network incorporates meticulously designed wavelet basis unit (WBU) functions into convolutional layer, creating the interpretable wavelet basis unit convolutional (WBUConv) layer. Convolutional kernels with clear physical significance enable the WBUConv layer to extract fault-related features in both time and frequency domains, enhancing diagnostic performance, and interpreting the CNN's attention frequency along with the convolutional kernel's training outcomes. In this paper, three WBU functions are designed to construct the corresponding WBUNs, and their effectiveness and interpretability are verified through three sets of mechanical fault diagnosis experiments. Meanwhile, experimental results demonstrate the WBUConv layer's remarkable advantages in noise robustness, convergence speed, and strong generalization ability.

show abstract

Multilayer Grad-CAM: An effective tool towards explainable deep neural networks for intelligent fault diagnosis

Cited by 34 publications

References 24 publications

Classification of LED Packages for Quality Control by Discriminant Analysis, Neural Network and Decision Tree

Classification of LED Packages for Quality Control by Discriminant Analysis, Neural Network and Decision Tree

A multi-modal joint attention network for vibro-acoustic fusion diagnosis of engines

WBUN: an interpretable convolutional neural network with wavelet basis unit embedded for fault diagnosis

Contact Info

Product

Resources

About