Reliability-Aware Support Vector Machine-Based High-Level Surrogate Model for Analog Circuits

Khandelwal, Sapna; Garg, Lokesh; Boolchandani, Dharmendar

doi:10.1109/tdmr.2015.2431436

Cited by 10 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This straight-forward approach treats the given circuit as a black box and uses some mathematical approximation, such as linear or non-linear regression, neural networks, Volterra series, Kriging technique, response surface design, etc., to build the relation equations between its inputs and outputs. These kinds of models, which are known as metamodels and surrogate models [12][13][14][15][16], require many simulation results to perform the approximating process. Due to extensive simulation time, it is impractical for large designs.…”

Section: Behavioral Modeling For Analog Circuitmentioning

confidence: 99%

“…If the required behavioral models can be automatically generated from the given designs, this could be an attractive approach for designers. Because the behavioral models of analog circuits and digital circuits are built in different ways [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28], it is necessary to correctly identify different circuit blocks in the transistor-level netlist of the whole system. This can also help to understand the behavior of the flattened transistor-level netlists extracted from layouts in reverse engineering or commercial tools.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Efficient Circuit Structure Analysis for Automatic Behavioral Model Generation in Mixed-Signal System Simulation

et al. 2021

View full text Add to dashboard Cite

For mixed-signal systems, identifying the analog and digital circuit blocks in the transistor-level netlist has many benefits for system analysis and verification. However, existing approaches still have difficulty handling large mixed-signal designs with millions of transistors, especially when multiple analog structure patterns are included. In this paper, we propose an efficient structure recognition methodology to support analyzing highly complex designs with various circuit structures and different devices. In order to tackle the complexity of real cases, a hierarchical partition-based analysis methodology and an encoding-based fast screening technique are proposed in this work. To correctly ascertain the boundary of analog and digital structures, we propose an enhanced direct current connection (DCC) partition method and combine it with the analog structure analysis flow. The non-transistor devices, such as resistors and capacitors, are also included in our recognition flow to improve the recognition capability and accuracy. As demonstrated with two industrial cases, the behavioral models generated from the structure recognition results do help to improve the efficiency of the AMS system verification.

show abstract

Section: Behavioral Modeling For Analog Circuitmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Efficient Circuit Structure Analysis for Automatic Behavioral Model Generation in Mixed-Signal System Simulation

et al. 2021

View full text Add to dashboard Cite

show abstract

“…e SVM is another well-known and effective supervised learning model for selecting attributes and classifying data. Before the rise of deep learning, the SVM outperformed ANNs in various real-life applications in medicine [3,4], semiconductor industry [18], online analysis [19], spectral unmixing resolution [20], imbalanced datasets [21], mining financial distress [22], data classification [23], and so forth [24][25][26]. In comparison with deep learning techniques that try to connect data in terms of ANNs, the SVM separates (not to connect) different classes of data based on the kernels through mathematical optimization [27,28].…”

Section: Introductionmentioning

confidence: 99%

A New Support Vector Machine Based on Convolution Product

et al. 2021

View full text Add to dashboard Cite

The support vector machine (SVM) and deep learning (e.g., convolutional neural networks (CNNs)) are the two most famous algorithms in small and big data, respectively. Nonetheless, smaller datasets may be very important, costly, and not easy to obtain in a short time. This paper proposes a novel convolutional SVM (CSVM) that has the advantages of both CNN and SVM to improve the accuracy and effectiveness of mining smaller datasets. The proposed CSVM adapts the convolution product from CNN to learn new information hidden deeply in the datasets. In addition, it uses a modified simplified swarm optimization (SSO) to help train the CSVM to update classifiers, and then the traditional SVM is implemented as the fitness for the SSO to estimate the accuracy. To evaluate the performance of the proposed CSVM, experiments were conducted to test five well-known benchmark databases for the classification problem. Numerical experiments compared favorably with those obtained using SVM, 3-layer artificial NN (ANN), and 4-layer ANN. The results of these experiments verify that the proposed CSVM with the proposed SSO can effectively increase classification accuracy.

show abstract

Research on Evaluation and Prediction Method of Electrical Life of Contactor Individual

Sun,

Liu,

Wang

et al. 2024

IEEJ Transactions Elec Engng

View full text Add to dashboard Cite

Aiming at the problem of electrical life prediction of AC contactor individual, considering that characteristic information and degradation trend of different life stages are different, a comprehensive electrical life analysis method is proposed. First, based on uncertainty reasoning, initial electrical life is qualitatively evaluated to realize an early warning for contactor with poor performance. Secondly, real‐time monitoring of performance degradation inflection point is used to realize the determination of degradation stage. With degradation stage matching mechanism, the quantitative prediction of remaining useful electrical life is realized based on performance inflection point and Elman neural network, thus forming an effective monitoring of the operation state in full life cycle. It is verified by experiment that the proposed method can accurately evaluate initial life grade and predict remaining useful life after performance inflection point. Compared with other methods, the information of electrical life for contactor individual can be accurately and comprehensively obtained. © 2024 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC.

show abstract

Reliability-Aware Support Vector Machine-Based High-Level Surrogate Model for Analog Circuits

Cited by 10 publications

References 8 publications

Efficient Circuit Structure Analysis for Automatic Behavioral Model Generation in Mixed-Signal System Simulation

Efficient Circuit Structure Analysis for Automatic Behavioral Model Generation in Mixed-Signal System Simulation

A New Support Vector Machine Based on Convolution Product

Research on Evaluation and Prediction Method of Electrical Life of Contactor Individual

Contact Info

Product

Resources

About