Data privacy preserving federated transfer learning in machinery fault diagnostics using prior distributions

2021 ◽  
pp. 147592172110292
Author(s):  
Wei Zhang ◽  
Xiang Li
Keyword(s):  
Transfer Learning ◽  
Data Privacy ◽  
Model Performance ◽  
Operating Conditions ◽  
Learning Performance ◽  
Fault Diagnostics ◽  
Prior Distributions ◽  
The Real ◽  
Diagnostic Problems ◽  
Fault Diagnostic

Federated learning has been receiving increasing attention in the recent years, which improves model performance with data privacy among different clients. The intelligent fault diagnostic problems can be largely benefited from this emerging technology since the private data generally cannot leave local storage in the real industries. While promising federated learning performance has been achieved in the literature, most studies assume data from different clients are independent and identically distributed. In the real industrial scenarios, due to variations in machines and operating conditions, the data distributions are generally different across different clients, that significantly deteriorates the performance of federated learning. To address this issue, a federated transfer learning method is proposed in this article for machinery fault diagnostics. Under the condition that data from different clients cannot be communicated, prior distributions are proposed to indirectly bridge the domain gap. In this way, client-invariant features can be extracted for diagnostics while the data privacy is preserved. Experiments on two rotating machinery datasets are implemented for validation, and the results suggest the proposed method offers an effective and promising approach for federated transfer learning in fault diagnostic problems.

scholarly journals A Study of Deep Neural Networks Transfer Learning For Fault Diagnosis Applications

2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Michael Franco-Garcia ◽  
Alex Benasutti ◽  
Larry Pearlstein ◽  
Mohammed Alabsi
Keyword(s):  
Neural Networks ◽  
Fault Diagnosis ◽  
Transfer Learning ◽  
Deep Neural Networks ◽  
Model Performance ◽  
Poor Performance ◽  
Operating Conditions ◽  
Learning Performance ◽  
Vibration Data ◽  
Model Training

Intelligent fault diagnosis utilizing deep learning algorithms has been widely investigated recently. Although previous results demonstrated excellent performance, features learned by Deep Neural Networks (DNN) are part of a large black box. Consequently, lack of understanding of underlying physical meanings embedded within the features can lead to poor performance when applied to different but related datasets i.e. transfer learning applications. This study will investigate the transfer learning performance of a Convolution Neural Network (CNN) considering 4 different operating conditions. Utilizing the Case Western Reserve University (CWRU) bearing dataset, the CNN will be trained to classify 12 classes. Each class represents a unique differentfault scenario with varying severity i.e. inner race fault of 0.007”, 0.014” diameter. Initially, zero load data will be utilized for model training and the model will be tuned until testing accuracy above 99% is obtained. The model performance will be evaluated by feeding vibration data collected when the load is varied to 1, 2 and 3 HP. Initial results indicated that the classification accuracy will degrade substantially. Hence, this paper will visualize convolution kernels in time and frequency domains and will investigate the influence of changing loads on fault characteristics, network classification mechanism and activation strength.

Power Test System Development and Dynamic Performance State Estimation Based on Hub Motor Vehicle

2020 ◽  
Vol 13 (2) ◽  
pp. 126-140
Author(s):  
Jing Gan ◽  
Xiaobin Fan ◽  
Zeng Song ◽  
Mingyue Zhang ◽  
Bin Zhao
Keyword(s):  
Real Time ◽  
Dynamic Performance ◽  
Test System ◽  
Operating Conditions ◽  
Motor Vehicles ◽  
Maximum Speed ◽  
Performance Parameters ◽  
Power Performance ◽  
Power Test ◽  
The Real

Background: The power performance of an electric vehicle is the basic parameter. Traditional test equipment, such as the expensive chassis dynamometer, not only increases the cost of testing but also makes it impossible to measure all the performance parameters of an electric vehicle. Objective: A set of convenient, efficient and sensitive power measurement system for electric vehicles is developed to obtain the real-time power changes of hub-motor vehicles under various operating conditions, and the dynamic performance parameters of hub-motor vehicles are obtained through the system. Methods: Firstly, a set of on-board power test system is developed by using virtual instrument (Lab- VIEW). This test system can obtain the power changes of hub-motor vehicles under various operating conditions in real-time and save data in real-time. Then, the driving resistance of hub-motor vehicles is analyzed, and the power performance of hub-motor vehicles is studied in depth. The power testing system is proposed to test the input power of both ends of the driving motor, and the chassis dynamometer is combined to test so that the output efficiency of the driving motor can be easily obtained without disassembly. Finally, this method is used to carry out the road test and obtain the vehicle dynamic performance parameters. Results: The real-time current, voltage and power, maximum power, acceleration time and maximum speed of the vehicle can be obtained accurately by using the power test system in the real road experiment. Conclusion: The maximum power required by the two motors reaches about 9KW, and it takes about 20 seconds to reach the maximum speed. The total power required to maintain the maximum speed is about 7.8kw, and the maximum speed is 62km/h. In this article, various patents have been discussed.

scholarly journals Effectiveness of transfer learning for enhancing tumor classification with a convolutional neural network on frozen sections

2020 ◽  
Vol 10 (1) ◽  
Author(s):  
Young-Gon Kim ◽  
Sungchul Kim ◽  
Cristina Eunbee Cho ◽  
In Hye Song ◽  
Hee Jin Lee ◽  
...  
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Convolutional Neural Network ◽  
Transfer Learning ◽  
Frozen Section ◽  
Medical Center ◽  
External Validation ◽  
Model Performance ◽  
Classification Model ◽  
Training Dataset

AbstractFast and accurate confirmation of metastasis on the frozen tissue section of intraoperative sentinel lymph node biopsy is an essential tool for critical surgical decisions. However, accurate diagnosis by pathologists is difficult within the time limitations. Training a robust and accurate deep learning model is also difficult owing to the limited number of frozen datasets with high quality labels. To overcome these issues, we validated the effectiveness of transfer learning from CAMELYON16 to improve performance of the convolutional neural network (CNN)-based classification model on our frozen dataset (N = 297) from Asan Medical Center (AMC). Among the 297 whole slide images (WSIs), 157 and 40 WSIs were used to train deep learning models with different dataset ratios at 2, 4, 8, 20, 40, and 100%. The remaining, i.e., 100 WSIs, were used to validate model performance in terms of patch- and slide-level classification. An additional 228 WSIs from Seoul National University Bundang Hospital (SNUBH) were used as an external validation. Three initial weights, i.e., scratch-based (random initialization), ImageNet-based, and CAMELYON16-based models were used to validate their effectiveness in external validation. In the patch-level classification results on the AMC dataset, CAMELYON16-based models trained with a small dataset (up to 40%, i.e., 62 WSIs) showed a significantly higher area under the curve (AUC) of 0.929 than those of the scratch- and ImageNet-based models at 0.897 and 0.919, respectively, while CAMELYON16-based and ImageNet-based models trained with 100% of the training dataset showed comparable AUCs at 0.944 and 0.943, respectively. For the external validation, CAMELYON16-based models showed higher AUCs than those of the scratch- and ImageNet-based models. Model performance for slide feasibility of the transfer learning to enhance model performance was validated in the case of frozen section datasets with limited numbers.

scholarly journals DeepFogSim: A Toolbox for Execution and Performance Evaluation of the Inference Phase of Conditional Deep Neural Networks with Early Exits Atop Distributed Fog Platforms

2021 ◽  
Vol 11 (1) ◽  
pp. 377
Author(s):  
Michele Scarpiniti ◽  
Enzo Baccarelli ◽  
Alireza Momenzadeh ◽  
Sima Sarv Ahrabi
Keyword(s):  
Neural Networks ◽  
Real Time ◽  
Future Internet ◽  
Operating Conditions ◽  
Graphic User Interface ◽  
Energy Aware ◽  
Internet Applications ◽  
The Real ◽  
Delay Performance ◽  
Hard Constraints

The recent introduction of the so-called Conditional Neural Networks (CDNNs) with multiple early exits, executed atop virtualized multi-tier Fog platforms, makes feasible the real-time and energy-efficient execution of analytics required by future Internet applications. However, until now, toolkits for the evaluation of energy-vs.-delay performance of the inference phase of CDNNs executed on such platforms, have not been available. Motivated by these considerations, in this contribution, we present DeepFogSim. It is a MATLAB-supported software toolbox aiming at testing the performance of virtualized technological platforms for the real-time distributed execution of the inference phase of CDNNs with early exits under IoT realms. The main peculiar features of the proposed DeepFogSim toolbox are that: (i) it allows the joint dynamic energy-aware optimization of the Fog-hosted computing-networking resources under hard constraints on the tolerated inference delays; (ii) it allows the repeatable and customizable simulation of the resulting energy-delay performance of the overall Fog execution platform; (iii) it allows the dynamic tracking of the performed resource allocation under time-varying operating conditions and/or failure events; and (iv) it is equipped with a user-friendly Graphic User Interface (GUI) that supports a number of graphic formats for data rendering. Some numerical results give evidence for about the actual capabilities of the proposed DeepFogSim toolbox.

Machine Learning as a Silent Observer of Advanced Geoscience Interpretation

2021 ◽  
Author(s):  
Muhammad Sajid
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Transfer Learning ◽  
Learning Performance ◽  
Learning Approach ◽  
Seismic Fault ◽  
Facies Classification ◽  
The World ◽  
Learning Machine ◽  
Fault Interpretation

Abstract Machine learning is proving its successes in all fields of life including medical, automotive, planning, engineering, etc. In the world of geoscience, ML showed impressive results in seismic fault interpretation, advance seismic attributes analysis, facies classification, and geobodies extraction such as channels, carbonates, and salt, etc. One of the challenges faced in geoscience is the availability of label data which is one of the most time-consuming requirements in supervised deep learning. In this paper, an advanced learning approach is proposed for geoscience where the machine observes the seismic interpretation activities and learns simultaneously as the interpretation progresses. Initial testing showed that through the proposed method along with transfer learning, machine learning performance is highly effective, and the machine accurately predicts features requiring minor post prediction filtering to be accepted as the optimal interpretation.

A federated interpretable scorecard and its application in credit scoring

2021 ◽  
pp. 2142009
Author(s):  
Fanglan Zheng ◽  
Erihe ◽  
Kun Li ◽  
Jiang Tian ◽  
Xiaojia Xiang
Keyword(s):  
Privacy Protection ◽  
Data Privacy ◽  
Credit Scoring ◽  
Training Session ◽  
Model Performance ◽  
Parameter Tuning ◽  
Kolmogorov Smirnov ◽  
Data Privacy Protection ◽  
Tuning Process ◽  
Positive Coefficients

In this paper, we propose a vertical federated learning (VFL) structure for logistic regression with bounded constraint for the traditional scorecard, namely FL-LRBC. Under the premise of data privacy protection, FL-LRBC enables multiple agencies to jointly obtain an optimized scorecard model in a single training session. It leads to the formation of scorecard model with positive coefficients to guarantee its desirable characteristics (e.g., interpretability and robustness), while the time-consuming parameter-tuning process can be avoided. Moreover, model performance in terms of both AUC and the Kolmogorov–Smirnov (KS) statistics is significantly improved by FL-LRBC, due to the feature enrichment in our algorithm architecture. Currently, FL-LRBC has already been applied to credit business in a China nation-wide financial holdings group.

Numerical Approach for Real Gas Simulations: Part II — Flow Simulation for Supercritical CO2 Centrifugal Compressor

2017 ◽  
Author(s):  
Swati Saxena ◽  
Ramakrishna Mallina ◽  
Francisco Moraga ◽  
Douglas Hofer
Keyword(s):  
Critical Point ◽  
Centrifugal Compressor ◽  
Supercritical Co2 ◽  
Operating Conditions ◽  
Inlet Guide Vane ◽  
Thermodynamic Quantities ◽  
3D Flow ◽  
Real Gas ◽  
The Real ◽  
Operating Range

This paper is presented in two parts. Part I (Tabular fluid properties for real gas analysis) describes an approach to creating a tabular representation of the equation of state that is applicable to any fluid. This approach is applied to generating an accurate and robust tabular representation of the RefProp CO2 properties. Part II (this paper) presents numerical simulations of a low flow coefficient supercritical CO2 centrifugal compressor developed for a closed loop power cycle. The real gas tables presented in part I are used in these simulations. Three operating conditions are simulated near the CO2 critical point: normal day (85 bar, 35C), hot day (105 bar, 50 C) and cold day (70 bar, 20C) conditions. The compressor is a single stage overhung design with shrouded impeller, 155 mm impeller tip diameter and a vaneless diffuser. An axial variable inlet guide vane (IGV) is used to control the incoming swirl into the impeller. An in-house three-dimensional computational fluid dynamics (CFD) solver named TACOMA is used with real gas tables for the steady flow simulations. The equilibrium thermodynamic modeling is used in this study. The real gas effects are important in the desired impeller operating range. It is observed that both the operating range (minimum and maximum volumetric flow rate) and the pressure ratio across the impeller are dependent on the inlet conditions. The compressor has nearly 25% higher operating range on a hot day as compared to the normal day conditions. A condensation region is observed near the impeller leading edge which grows as the compressor operating point moves towards choke. The impeller chokes near the mid-chord due to lower speed of sound in the liquid-vapor region resulting in a sharp drop near the choke side of the speedline. This behavior is explained by analyzing the 3D flow field within the impeller and thermodynamic quantities along the streamline. The 3D flow analysis for the flow near the critical point provides useful insight for the designers to modify the current compressor design for higher efficiency.

scholarly journals Transfer Learning with Deep Recurrent Neural Networks for Remaining Useful Life Estimation

2018 ◽  
Vol 8 (12) ◽  
pp. 2416 ◽  
Author(s):  
Ansi Zhang ◽  
Honglei Wang ◽  
Shaobo Li ◽  
Yuxin Cui ◽  
Zhonghao Liu ◽  
...  
Keyword(s):  
Neural Networks ◽  
Transfer Learning ◽  
Recurrent Neural Networks ◽  
Prediction Models ◽  
Learning Algorithm ◽  
Remaining Useful Life ◽  
Operating Conditions ◽  
Training Data ◽  
Good Prediction ◽  
Useful Life

Prognostics, such as remaining useful life (RUL) prediction, is a crucial task in condition-based maintenance. A major challenge in data-driven prognostics is the difficulty of obtaining a sufficient number of samples of failure progression. However, for traditional machine learning methods and deep neural networks, enough training data is a prerequisite to train good prediction models. In this work, we proposed a transfer learning algorithm based on Bi-directional Long Short-Term Memory (BLSTM) recurrent neural networks for RUL estimation, in which the models can be first trained on different but related datasets and then fine-tuned by the target dataset. Extensive experimental results show that transfer learning can in general improve the prediction models on the dataset with a small number of samples. There is one exception that when transferring from multi-type operating conditions to single operating conditions, transfer learning led to a worse result.

Sign in / Sign up

Export Citation Format

Share Document