Learning from Limited Data in VLSI CAD
Applying machine learning to analyze data from design and test flows has received growing interests in recent years. In some applications, data can be limited and the core of analytics becomes a feature search problem. In this context, the chapter explains the challenges for adopting a traditional machine learning problem formulation view. An adjusted machine learning view is suggested where learning from limited data is treated as an iterative feature search process. The theoretical and practical considerations for implementing such a search process are discussed.
This work is supported in part by National Science Foundation under grant No. 1618118 and in part by Semiconductor Research Corporation with project 2016-CT-2706.
This chapter is an extension from author’s prior work “Machine Learning for Feature-Based Analytics,” in Proceeding 2018 International Symposium on Physical Design, pp. 74-81 Ⓒ2018 Association for Computing Machinery, Inc. http://doi.acm.org/10.1145/3177540.3177555. Reprinted by permission.
- 2.W. Chen, L.-C. Wang, J. Bhadra, Simulation knowledge extraction and reuse in constrained random processor verification, in ACM/IEEE Design Automation Conference (2013)Google Scholar
- 3.L.-C. Wang, Data mining in functional test content optimization, in ACM/IEEE Asian South Pacific Design Automation Conference (2015)Google Scholar
- 4.G. Drmanac, F. Liu, L.-C. Wang, Predicting variability in nanoscale lithography processes, in ACM/IEEE Design Automation Conference (2009)Google Scholar
- 5.J. Chen, B. Bolin, L.-C. Wang, J. Zeng, D. (Gagi) Drmanac, M. Mateja, Mining AC delay measurements for understanding speed-limiting paths, in IEEE International Test Conference (2010)Google Scholar
- 7.J. Chen, L.-C. Wang, P.-H. Chang, J. Zeng, S. Yu, M. Metaja, Data learning techniques and methodology for Fmax prediction, in IEEE International Test Conference (2009)Google Scholar
- 8.J. Tikkanen, S. Siatkowski, N. Sumikawa, L.-C. Wang, M.S. Abadir, Yield optimization using advanced statistical correlation methods, in IEEE International Test Conference (2014)Google Scholar
- 9.S. Siatkowski, L.-C. Wang, N. Sumikawa, L. Winemberg, Learning the process for correlation analysis, in IEEE VLSI Test Symposium (2017)Google Scholar
- 11.N. Callegari, D. (Gagi) Drmanac, L.-C. Wang, M.S. Abadir, Classification rule learning using subgroup discovery of cross-domain attributes responsible for design-silicon mismatch, in ACM/IEEE Design Automation Conference, 374–379 (2010)Google Scholar
- 13.M. Nero, J. Shan, L.-C. Wang, N. Sumikawa, Concept recognition in production yield data analytics, in IEEE International Test Conference (2018)Google Scholar
- 14.K. Hsieh, L.-C. Wang, W. Chen, J. Bhadra, Learning to produce direct tests for security verification using constrained process discovery, in Design Automation Conference (2017)Google Scholar
- 15.K-K. Hsieh, S. Siatkowski, L.-C. Wang, W. Chen, J. Bhadra, Feature extraction from design documents to enable rule learning for improving assertion coverage, in ACM/IEEE Asia South Pacific Design Automation Conference (2017)Google Scholar
- 20.A. Daniely, N. Linial, S. Shaleve-Shwartz, From average case complexity to improper learning complexity, in ACM Symposium on Theory of Computing (2014), pp. 441–448.Google Scholar
- 23.A. Daniely, S. Shalev-Shwartz, Complexity theoretic limitations on learning DNF’s, in PMLR: Proceeding of Machine Learning Research, vol. 49 (2016), pp. 815–830Google Scholar
- 28.K. Hsieh, L.-C. Wang, A concept learning tool based on calculating version space cardinality (2018). arXiv:1803.08625v1Google Scholar
- 29.A. Biere, Lingeling, Plingeling and Treengeling entering the SAT competition, in Proceedings of SAT Competition (2013)Google Scholar
- 30.I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, J. Bengio, Generative adversarial networks (2014). arXiv:1406.2661Google Scholar
- 31.A. Radford, L. Metz, S. Chintala, Unsupervised representation learning with deep convolutional generative adversarial networks (2016). arXiv:1511.06434v2Google Scholar