Adaptive Learning Based Importance Sampling for Analog Circuit DPPM Estimation

Yılmaz, Emre; Ozev, Sule

doi:10.1109/mdat.2014.2361719

Cited by 7 publications

(3 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Extreme value theory is used in [38] to estimate the performance of the test in the situation where the test boundaries lie in the tails of the distribution. Other techniques such as importance sampling [39] and statistical blockade [40]- [43] rely on altering the selection likelihood of Monte Carlo samples in such a way that sampling is pushed towards the test boundaries. This way we guarantee that the model is trained with both passing and failing devices, mimicking what should be observed in the production line.…”

Section: Machine Learning Indirect Testmentioning

confidence: 99%

Mixed-signal test automation: Are we there yet?

Léger

Barragán

2018

2018 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

Testing analog, mixed-signal and RF (AMS-RF) circuits represents a significant cost component for testing complex SoCs. Moreover, AMS-RF test generation and validation are still largely handcrafted tasks that rely on expert design knowledge for each particular Device Under Test (DUT). Mixed-signal test automation has been sought by the test community for the last decades, trying to mimic the success of digital test approaches. Indeed, in the digital domain, test is vastly automated and standard techniques are already available (ATPGs, BIST, scan registers, etc.). In the last decade, a methodology based on leveraging the power of machine learning algorithms has been proposed for AMS-RF circuits that opens the door to a higher level of automation. In this paper we review recent results in this line and try to put together what could be such a complete methodology and what remains to be done.

show abstract

Section: Machine Learning Indirect Testmentioning

confidence: 99%

Mixed-signal test automation: Are we there yet?

Léger

Barragán

2018

2018 IEEE International Symposium on Circuits and Systems (ISCAS)

View full text Add to dashboard Cite

show abstract

“…In the context of parametric test metrics evaluation, several fast Monte Carlo alternative approaches have been proposed to date. Approaches based on density estimation [5], Copulas theory [6], extreme value theory [7], importance sampling [8], and generation of parametric fault models [9], [10] require that the circuit can be simulated at transistor-level at least a few hundred or thousand times (e.g. amplifiers, filters, mixers, bandgap reference, etc.…”

Section: Introductionmentioning

confidence: 99%

Practical Simulation Flow for Evaluating Analog/Mixed-Signal Test Techniques

et al. 2016

View full text Add to dashboard Cite

Alternative test techniques are continuously proposed for analog and mixed-signal circuits with the aim to reduce the standard test cost and complexity. One of the main reasons why the majority of these alternative test solutions have not been met with success is the lack of a proof that they will not sacrifice the accuracy of the standard test, resulting in intolerable test escapes and yield loss. In this paper, we target specifically circuits with long simulation times and we present a practical simulation flow that allows evaluating in a time-efficient manner alternative test solutions during their development phase by estimating the resultant parametric test metrics. The simulation flow can be used to provide feedback for refining an alternative test solution and to give the green light for implementing the alternative test solution in high-volume production. The simulation flow is demonstrated for evaluating a built-in self-test strategy for ΣΔ ADCs.

show abstract

“…Several fast alternatives to Monte Carlo simulation exist [10]. In the context of parametric test metrics evaluation, several approaches have been studied, including density estimation [11], [12], Copulas theory [13], extreme value theory [14], [15], importance sampling [16], and generation of parametric fault models [17], [18]. However, all these approaches make the tacit assumption that the circuit can be simulated at transistor-level at least a few hundreds or thousand times.…”

Section: Introductionmentioning

confidence: 99%