Learning from Production Test Data: Correlation Exploration and Feature Engineering

Lin, Fan; Hsu, C. S.; Cheng, Kwang-Ting

doi:10.1109/ats.2014.51

Cited by 9 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, an evaluation if those methods could be used with datasets containing numerical and non-numerical variables is included. On the other hand, missing data was evaluated only in reference [3] where the case of stop-on fail scenario was covered. A data pre-processing is needed to handle missing values when using any of the other methods.…”

Section: A Qualitative Comparison Of Techniquesmentioning

confidence: 99%

“…A data pre-processing is needed to handle missing values when using any of the other methods. Different methods for linear correlation analysis have been used for testing reduction [1][2][3]. However, these methods cannot be used for non-parametric variables.…”

Section: A Qualitative Comparison Of Techniquesmentioning

confidence: 99%

“…Parametric analyses are widely used to predict the outcome of a test set, for instance in [1] the authors analysed correlation between each pair of tests items. Variations of Group LASSO method were used in [2] and [3] to identify tests which could be predicted by a linear combination of other tests. Furthermore, the last one covers stop-to-fail scenario where the test program stops as soon as a device fails a test and the remaining tests will not be applied Authors of work published in [5] applied Chi-Square to discard test with very small significance values and then used Support Vector Machine (SVM), in particular C-Support Vector Classification (C-SVC), to carry out the test process compaction.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Data Driven Predictive Model to Compact a Production Stop-on-Fail Test Set for an Electronic Device

Hinojosa

Stoyanov

2018

2018 International Conference on Computing, Electronics &Amp; Communications Engineering (iCCECE)

View full text Add to dashboard Cite

Decision Tree is a popular machine learning algorithm used for fault detection and classification in the industry. In this paper, the modelling technique is used to compact a production test set defined for quality assurance of an electronic asset. The novelty of this work is in the proposed method that builds in an iterative way decision trees until an accurate predictive model that meets classification accuracy target in a stop-on-fail test scenario. Generated test data is characterized with missing values which is a major challenge to the traditional use of decision trees. The developed computational procedure handles this application-specific data attribute. Exemplary results show that the method is able to significantly reduce a production test set with parametric and non-parametric tests, and generate a truthful prognostic model. In addition, the method is computationally efficient and easy to implement. It could also be combined with another test compaction strategies such as variables association analysis. Furthermore, the method proposed offers the flexibility of exploring the trade-off between the number of removed tests from the production test set and the prediction accuracy. The results can enable production costs reduction without impacting quality detection accuracy. The paper details and provides discussions on the advantages and limitations of the proposed algorithm.

show abstract

Section: A Qualitative Comparison Of Techniquesmentioning

confidence: 99%

Section: A Qualitative Comparison Of Techniquesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data Driven Predictive Model to Compact a Production Stop-on-Fail Test Set for an Electronic Device

Hinojosa

Stoyanov

2018

2018 International Conference on Computing, Electronics &Amp; Communications Engineering (iCCECE)

View full text Add to dashboard Cite

show abstract

“…Machine learning has wide application in circuit testing [13]. There are three main categories: (i) outlier identification [21], (ii) test cost reduction [22, 23], (iii) feature ranking [10], and (iv) circuit performance prediction [10, 24]. In particular, machine learning considering voltage droop has been proposed in [12].…”

Section: Past Researchmentioning

confidence: 99%

PSN‐aware circuit test timing prediction using machine learning

Liu

Han

Lin

et al. 2017

IET Computers & Digital Techniques

View full text Add to dashboard Cite

Excessive power supply noise (PSN) such as IR drop can cause yield loss when testing very large scale integration chips. However, simulation of circuit timing with PSN is not an easy task. In this study, the authors predict circuit timing for all test patterns using three machine learning techniques, neural network (NN), support vector regression (SVR), and least-square boosting (LSBoost). To reduce the huge dimension of raw data, they propose four feature extractions: input/output transition (IOT), flip-flop transition in window (FFTW), switching activity in window (SAW), and terminal FF transition of long paths (PATH). SAW and FFTW are physical-aware features while PATH is a timing-aware feature. Their experimental results on leon3mp benchmark circuit (638 K gates, 2 K test patterns) show that, compared with the simple IOT method, SAW effectively reduced the dimension by up to 472 times, without significant impact on prediction accuracy [correlation coefficient = 0.79]. Their results show that NN has best prediction accuracy and SVR has the least under-prediction. LSBoost uses the least memory. The proposed method is more than six orders of magnitude faster than traditional circuit simulation tools.

show abstract

An artificial neural network approach for screening test escapes

Lin

Cheng

2017

2017 22nd Asia and South Pacific Design Automation Conference (ASP-DAC)

View full text Add to dashboard Cite

Learning from Production Test Data: Correlation Exploration and Feature Engineering

Cited by 9 publications

References 36 publications

Data Driven Predictive Model to Compact a Production Stop-on-Fail Test Set for an Electronic Device

Data Driven Predictive Model to Compact a Production Stop-on-Fail Test Set for an Electronic Device

PSN‐aware circuit test timing prediction using machine learning

An artificial neural network approach for screening test escapes

Contact Info

Product

Resources

About