Building partially understandable convolutional neural networks by differentiating class-related neural nodes

Dai, Dawei; Tang, Chengfu; Wang, Guoyin; Xia, Shuyin

doi:10.1016/j.neucom.2021.04.003

Cited by 7 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We need to observe the essence through the phenomenon, so as to explain the predicted results. As mentioned in the paper [48], the CNN model is like a black box. How to enhance the interpretability of the model is very necessary, especially in the field of rolling bearing fault diagnosis.…”

Section: Interpretability Of Lightweight 1dcnnmentioning

confidence: 99%

Explainable 1DCNN with demodulated frequency features method for fault diagnosis of rolling bearing under time-varying speed conditions

Lu¹,

Tong²,

Feng³

et al. 2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

Intelligent fault diagnosis of rolling bearings under non-stationary and time-varying speed conditions is still a challenging task. At the same time, a reasonable explanation for an intelligent diagnosis model based on features is currently lacking. Therefore, we exploit an explainable one-dimensional convolutional neural network (1DCNN) model by combining with the demodulated frequency features of vibration signals and apply it to the fault classification of rolling bearings under time-varying speed conditions. First, the speed signals obtained by the speed encoder were transformed into generalized demodulation operator (GDO). Second, combined with the sensitive frequency band and GDO, the generalized demodulation algorithm was used to extract the frequency features from the amplitude envelope of the vibration signal. Subsequently, the proposed lightweight 1DCNN was trained to classify the frequency features and identify the health states of the rolling bearing. Finally, the local interpretable model-agnostic explanations model was utilized to explain the proposed model based on the features which own weight. It is found that the internal classification mechanism of the lightweight 1DCNN is realized according to the distribution of fault features, which is consistent with the process of human brain analysis. Two kinds of time-varying speed datasets which come from the University of Ottawa and XJTU are tested and verified. The results show that compared with other intelligent fault diagnosis methods, the identification error of the proposed method is lower and the diagnosis stability is better. The average diagnostic accuracy was 96.26% and 99.82%, respectively.

show abstract

Section: Interpretability Of Lightweight 1dcnnmentioning

confidence: 99%

Explainable 1DCNN with demodulated frequency features method for fault diagnosis of rolling bearing under time-varying speed conditions

Lu¹,

Tong²,

Feng³

et al. 2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Dai et al built the business model from the perspective of business ecology. The elements of the model included target customers, business systems, strategic positioning, and partners [ 9 ]. The research on business model innovation is mainly on the innovation driving force, including the decision-making ability of the executives and the overall level of the enterprise technology.…”

Section: Review and Analysis Of Related Researchmentioning

confidence: 99%

Construction of Internet Business Innovation Network Model and Data Image Research under the Background of Game Theory

Bai

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Facing the fierce market competition, enterprises not only need a deeper understanding and concept renewal in marketing theory but also need a set of standardized, practical, and efficient technical means and methods. Therefore, it is urgent to develop advanced marketing analysis tools and marketing decision-making methods. Based on game theory and neural network model, this study simplifies the existing research methods. By introducing different models, an image analysis and decision-making model based on game theory and neural network is constructed. It mainly aims at enterprise decision-making. The method used is to simulate various decision-making processes of enterprises by establishing neural network model and game model. At the same time, the image simulation of the model is carried out. The results show that the highest market share of the selected products is 36.1%, and the highest brand awareness is 9 points. The product with the second market share has better quality, 8 brand awareness points, and the highest dealer fee (2.3 yuan). Market share is less affected by product price and dealer expenses. The accuracy of the designed market share neural network model is 93%. This shows that the increase of market share is not realized simply by reducing the price but by increasing the profits of distributors and improving the brand image. Market leaders have the greatest revenue and profits. There is a positive correlation between the efforts of managers and the results achieved. Internet employees’ work effort is positively correlated with their basic salary. Different decisions have different effects on business models. The research of this paper provides a new idea for the innovation and development of Internet enterprise business model.

show abstract

Deep Transfer Learning in Mechanical Intelligent Fault Diagnosis: Application and Challenge

Qian

Zhu

Shen

et al. 2022

Neural Process Lett

View full text Add to dashboard Cite

Building partially understandable convolutional neural networks by differentiating class-related neural nodes

Cited by 7 publications

References 12 publications

Explainable 1DCNN with demodulated frequency features method for fault diagnosis of rolling bearing under time-varying speed conditions

Explainable 1DCNN with demodulated frequency features method for fault diagnosis of rolling bearing under time-varying speed conditions

Construction of Internet Business Innovation Network Model and Data Image Research under the Background of Game Theory

Deep Transfer Learning in Mechanical Intelligent Fault Diagnosis: Application and Challenge

Contact Info

Product

Resources

About