臺灣博碩士論文加值系統

隨著CMOS技術的不斷縮小，當今的集成電路（IC）可以提供更複雜的功能，並且能更廣泛的應用於各種應用中。其中一種應用是工業自動化。有了可靠的工業自動化系統可以因此提高生產率，降低人工成本並為公司帶來更好的利潤。但是到了現代,自動化生產系統在製造過程中仍然會遭遇到突發性的故障,最主要的威脅之一就是電路的老化,其中電路老化中又以負偏壓溫度不穩定性（NBTI）對現代IC構成了最嚴重威脅，並可能導致時序和功能故障。如果這些故障發生在工業自動化生產系統上，則由於製造質量和成品率不可接受，故障系統可能會造成重大的經濟損失。儘管預防性維護是避免此類情況的有用方法，但是頻繁執行預防性維護也會給生產線帶來大量停機時間。為了在即將發生電路故障之前準確地執行預防性維護，需要晶片剩餘壽命估計方法。在本文中，我們提出了一個預測晶片剩餘壽命的框架，該框架可以適應工藝和工作電壓的變化。該框架通過機器學習方法追蹤代表性的老化指標，以預測晶片的剩餘壽命。此外，我們還研究了諸如訓練樣本大小之類的超參數變化對預測性能的影響。實驗結果表示，我們的框架實現了97.3％和97.2％的平均準確度和精確度，與之前的工作相比，我們的準確度還要高出了2.54％。

As CMOS technology continues scaling down, integrated circuits (ICs) nowadays can provide more complex functions and are widely used in various applications. One of these applications is industrial automation. With reliable industrial automation systems, the industrial automation can increase the productivity, reduce labor cost, and make better profit for a company. However, the industrial automated production systems nowadays may suffer from unexpected failures during the fabrication process. One of the major threats of the system failure comes from the circuit aging, and in the aging of the circuit, the Negative-Bias Temperature Instability (NBTI) poses the most serious threats to modern ICs and may lead to timing and functional failure. If these failures happen at industrial automated production systems, the malfunctioning system can cause significant economic losses due to unacceptable fabrication quality and yield. Although preventive maintenance is a useful way to avoid such a situation, frequently executing preventive maintenance will also introduce significant downtime to the production line. To accurately execute the preventive maintenance just before circuit failure occurs, a chip remaining lifetime estimation method is in demand. In this thesis, we propose a framework for predicting the remaining lifetime of the chip, which can adapt to changes in the process and operating voltage. The framework tracks representative aging indicators through machine learning methods in order to predict the remaining lifetime of the chip. In addition, we also investigate the impact of changes of hyperparameters, such as training sample sizes, on prediction performance. Experimental results show that our framework achieves an average accuracy and precision of 97.3% and 97.2%, and our accuracy is 2.54% higher than the strategy of chip health level compared to a previous work.

摘要 iii
Abstract v
誌謝 vii
Table of Contents viii
List of Figures x
Chapter 1. Introduction 1
1.1 Background 1
1.2 Main Contribution 4
Chapter 2. Preliminaries 6
2.1 NBTI and NBTI Model 6
2.2 Timing Margin Detector 7
2.3 Chip Health Estimation Indicators 9
2.4 Machine Learning Model 11
Chapter 3. Proposed Framework 17
3.1 Aging Dataset Pre-Processing 18
3.2 Hyperparameter-Tuning 21
3.3 Model-Training 22
3.4 Ensemble Learning 23
Chapter 4. Evaluation 25
4.1 Experimental Setup 25
4.2 Evaluation Metrics 26
4.3 Experimental Result 27
Chapter 5. Conclusion and Future Work 32
References 33

[1] Predictive Technology Model (PTM). [Online]. Available: http://www.ptm.asu.edu. [Accessed: Dec. 4, 2020].
[2] Random Forest Simple Explanation. [Online]. Available: https://medium.com/@williamkoehrsen/random-forest-simple-explanation-377895a60d2d. [Accessed: Dec. 4, 2020].
[3] M.A. Alam, S. Mahapatra, A comprehensive model of PMOS NBTI degradation, In Microelectronics Reliability, pages 45:71-81, 2005.
[4] S. Arlot and A. Celisse. A survey of cross-validation procedures for model selection. In Statistics Surveys, vol. 4, pages 40–79, Mar 2010.
[5] L. Bottou. Stochastic gradient descent tricks. In Neural Networks: Tricks of the trade. Springer, pages 421–436, 2012.
[6] L. Breiman. Random forests. In Machine Learning, 45(1):5–32, Oct 2001.
[7] N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer. Smote: Synthetic minority over-sampling technique. In Journal of Artificial Intelligence Research, vol. 16, no. 1, pages 321–357, Jun 2002.
[8] Y.-G. Chen, I.-C. Lin, J.-T. Ke. ROAD: Improving reliability of multi-core system via asymmetric aging. In IEEE/ACM International Conference on Computer-Aided Design (ICCAD), 2019.
[9] J.-H. Chiang and J. Yuan. Optimal maintenance policy for a production system subject to aging and shocks. In Journal of Quality in Maintenance Engineering, vol. 6 No. 3, pages 200-216, 2000.
[10] C. Cortes and V. Vapnik. Support-vector networks. In Machine Learning, vol. 20, no. 3, pages 273–297, 1995.
[11] F. Firouzi, F. Ye, K. Chakrabarty, and M. B. Tahoori. Representative critical-path selection for aging-induced delay monitoring. In International Test Conference, Anaheim, CA, USA, pages 1–10, 2013.
[12] K.A. Grajski, L. Breiman, G.V. Di Prisco, and W.J. Freeman. Classification of EEG spatial patterns with a tree-structured methodology: CART. In IEEE Transactions on Biomedical Engineering (T-BME), pages 1076–1086, 1986.
[13] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, Jun 2016.
[14] B. Hea, Y. Shi, Q. Wan, and X. Zhao. Prediction of customer attrition of commercial banks based on SVM model. In 2nd International Conference on Information Technology and Quantitative Management (ITQM 2014), pages 423– 430, 2014.
[15] S. Huang, M. Aydin and T. A. Lipo. TORUS concept machines: preprototyping design assessment for two major topologies. In Conference Record of the 2001 IEEE Industry Applications Conference. 36th IAS Annual Meeting (Cat. No.01CH37248), Chicago, IL, USA, pages 1619-1625, 2001.
[16] K. Kang, K. Kim, A. E. Islam, M. A. Alam, and K. Roy. Characterization and estimation of circuit reliability degradation under NBTI using on-line I_DDQ measurement. In ACM/IEEE Design Automation Conference (DAC), pages 358–363, 2007.
[17] L. Lai, V. Chandra, R. C. Aitken, and P. Gupta. SlackProbe: A Flexible and efficient in situ timing slack monitoring methodology. In IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 33 no. 8, pages 1168–1179, Aug 2014.
[18] S. Mahapatra, N. Goel, S. Desai, S. Gupta, B. Jose, S. Mukhopadhyay, K. Joshi, A. Jain, A.E. Islam and M.A. Alam. A Comparative Study of Different Physics-Based NBTI Models. In IEEE Transactions on Electron Devices, vol. PP, no. 99, pages 1 –16, 2013.
[19] E. Mintarno, J. Skaf, R. Zheng, J. Velamala, Y. Cao, S. Boyd, R.W. Dutton, and S. Mitra. Optimized self-tuning for circuit aging. In Design, Automation and Test in Europe Conference (DATE), pages 586–591, 2010.
[20] S. Mishra, Narendra Parihar, Rakesh Rao, and Souvik Mahapatra, IIT Bombay, Mumbai, India Hiu Yung Wong, Steve Motzny, and Victor Moroz, Synopsys Inc., Mountain View. NBTI Modeling in Sentaurus Device. In TCAD News, pages 7-8, Dec 2016. [Online]. Available: https://www.synopsys.com/content/dam/synopsys/silicon/newsletters/tcad-news-winter2016-nl.pdf. [Accessed: Dec. 4, 2020].
[21] B. C. Paul, K. Kang, H. Kufluoglu, M. A. Alam, and K. Roy. Temporal performance degradation under NBTI: Estimation and design for improved reliability of nanoscale circuits. In ACM/IEEE Design, Automation and Test in Europe (DATE), pages 780–785, 2006.
[22] R. Peng, H. Xiao, and H. Liu. Reliability of multi-state systems with a performance sharing group of limited size. In Reliability Engineering and System Safety, vol. 166, pages 164–170, 2017.
[23] L. Prechelt. Early stopping – but when? In Neural Networks: Tricks of the Trade. Springer, pages 53–67, 2012.
[24] B. Qian and K. Rasheed. Foreign Exchange Market Prediction with Multiple Classifiers. In Journal of Forecasting, vol. 29, pages 271–284, 2010.
[25] J.J. Rodriguez, L.I. Kuncheva, and C.J. Alonso. Rotation forest: A new classifier ensemble method. In The IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), vol. 28, pages 1619-1630, Oct 2006.
[26] A. Sachdeva, D. Kumar, and P. Kumar. Planning and optimizing the maintenance of paper production systems in a paper plant. In Computers & Industrial Engineering, vol. 55, no. 4, pages 817–829, Nov 2008.
[27] T.A. Singh and S. Namrata. Speech Recognition Using Eulidean Distance. In International Journal of Emerging Technology and Advanced Engineering, vol. 3, no. 3, pages 587-590, 2013.
[28] S. Suthaharan. Machine learning models and algorithms for big data classification. In Integrated Series in Information Systems, vol. 36, New York, NY, USA: Springer, 2016.
[29] R. Vattikonda, W. Wang, and Y. Cao. Modeling and minimization of PMOS NBTI effect for robust nanometer design. In ACM/IEEE Design Automation Conference (DAC), pages 1047–1052, Jul 2006.
[30] W. Wang, Z. Wei, S. Yang, and Y. Cao. An efficient method to identify critical gates under circuit aging. In IEEE/ACM International Conference on Computer-Aided Design (ICCAD), pages 735–740, Nov 2007.
[31] X. Wang, K. Nguyen, and B. P. Nguyen. Churn Prediction using Ensemble Learning. In The 4th International Conference on Machine Learning and Soft Computing (ICMLSC 2020), January 17–19, 2020.
[32] Y.-T. Wang, K.-C. Wu, C.-H. Chou, and S.-C. Chang. Aging-aware chip health prediction adopting an innovative monitoring strategy. In Asia and South Pacific Design Automation Conference (ASP-DAC), pages 179–184, 2019.
[33] J. Wen, S. Pan, N. Chang, W.-T. Chuang, W. Xia, D. Zhu, A. Kumar, E.-C. Yang, K. Srinivasan, and Y. Li. Dnn-based fast static on-chip thermal solver. In Semiconductor Thermal Measurement, Modeling & Management Symposium (SEMI-THERM), 2020.
[34] T.-Y. Wu, S. Gharahi, and J. A. Abraham. An area efficient on-chip static IR drop detector/evaluator. In IEEE International Symposium on Circuits and Systems (ISCAS), pages 2009–2012, May 2009.
[35] L. Yang, Y. Zhao, R. Peng, and X. B. Ma. Opportunistic maintenance of production systems subject to random wait time and multiple control limits. In Journal of Manufacturing Systems, vol. 47, pages 12–34, Apr 2018.
[36] M.-L. Zhang and Z.-H. Zhou. A k-nearest neighbor based algorithm for multi-label classification. In IEEE International Conference on Granular Computing (GRC), pages 718–721, Jul 2005.
[37] X. Zhang, N. Tuzzio, and M. Tehranipoor . Identification of Recovered ICs using Fingerprints from a Light-Weight On-Chip Sensor. In ACM/IEEE Design Automation Conference (DAC), 2012.