Extraction of piecewise-linear analog circuit models from trained neural networks using hidden neuron clustering

  title={Extraction of piecewise-linear analog circuit models from trained neural networks using hidden neuron clustering},
  author={Simona Doboli and Gaurav Gothoskar and Alex Doboli},
  journal={2003 Design, Automation and Test in Europe Conference and Exhibition},
This paper presents a new technique for automatically creating analog circuit models. The method extracts - from trained neural networks-piecewise linear models expressing the linear dependencies between circuit performances and design parameters. The paper illustrates the technique for an OTA circuit for which models for gain and bandwidth were automatically generated. The extracted models have a simple form that accurately fits the sampled points and the behavior of the trained neural… 

Figures from this paper

Accurate Performance Estimation using Circuit Matrix Models in Analog Circuit Synthesis

Techniques such as hashing and near neighbor searches are proposed to expedite the matrix model evaluation procedure and give accurate results for synthesis for various types of circuit specifications.

Fast time-domain symbolic simulation for synthesis of sigma-delta analog-digital converters

  • Hui ZhangA. Doboli
  • Computer Science
    2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512)
  • 2004
This work presents a fast time-domain simulation method for automated synthesis of /spl Sigma/-/spl Delta/ ADC that uses a symbolic description of the ADC netlist and uses more complex circuit macromodels.

Accurate Analog Synthesis Based On Circuit Matrix Models

This dissertation proposes performance prediction using Circuit Matrix Models to estimate the performance of analog circuits and develops techniques to expedite synthesis by making matrix model evaluation extremely fast.

On the Use of Hash Tables for Efficient Analog Circuit Synthesis

Experimental results show that the proposed method can provide 4x-6x speedup over that offered by synthesis approaches employing macromodels but no hashing, thus improving synthesis time.



Piecewise-linear modeling of analog circuits based on model extraction from trained neural networks

  • G. GothoskarA. DoboliS. Doboli
  • Computer Science
    Proceedings of the 2002 IEEE International Workshop on Behavioral Modeling and Simulation, 2002. BMAS 2002.
  • 2002
A new technique for automatically creating analog circuit models is presented, which extracts piecewise linear models from trained neural networks and has simple form that accurately fits the sampled points.

Extraction of rules from artificial neural networks for nonlinear regression

An approach for extracting rules from trained NNs for regression by identifying a subregion of the input space and a linear function involving the relevant input attributes of the data approximates the network output for all data samples in this subregion.

Piecewise Linear Modeling and Analysis

Piecewise Linear Modeling and Analysis shows in detail how many existing components in electrical networks can be modeled and ranked to compare them and to show which model can handle the largest class of PL mappings.

Computer-aided design of analog and mixed-signal integrated circuits

This survey presents an overview of recent advances in the state of the art for computer-aided design (CAD) tools for analog and mixed-signal integrated circuits (ICs) and outlines progress on the various design problems involved.

Pattern classification and scene analysis

The topics treated include Bayesian decision theory, supervised and unsupervised learning, nonparametric techniques, discriminant analysis, clustering, preprosessing of pictorial data, spatial filtering, shape description techniques, perspective transformations, projective invariants, linguistic procedures, and artificial intelligence techniques for scene analysis.


  • van Bokhoven, “Piecewise Linear Modeling and Analysis”, Kluwer
  • 1998