Automated Network Design-The Frequency-Domain Case

Director, Stephen W.; Rohrer, R.A.

doi:10.1109/tct.1969.1082967

Cited by 187 publications

(35 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The minimisation of errors through gradient descent (Hadamard, 1908) in the parameter space of complex, nonlinear, differentiable (Leibniz, 1684), multi-stage, NN-related systems has been discussed at least since the early 1960s (e.g., Kelley, 1960;Bryson, 1961;Bryson and Denham, 1961;Pontryagin et al, 1961;Dreyfus, 1962;Wilkinson, 1965;Amari, 1967;Bryson and Ho, 1969;Director and Rohrer, 1969), initially within the framework of Euler-LaGrange equations in the Calculus of Variations (e.g., Euler, 1744).…”

Section: -1981 and Beyond: Development Of Backpropagation (Bp) Fomentioning

confidence: 99%

Deep learning in neural networks: An overview

2015

View full text Add to dashboard Cite

In recent years, deep artificial neural networks (including recurrent ones) have won numerous contests in pattern recognition and machine learning. This historical survey compactly summarises relevant work, much of it from the previous millennium. Shallow and deep learners are distinguished by the depth of their credit assignment paths, which are chains of possibly learnable, causal links between actions and effects. I review deep supervised learning (also recapitulating the history of backpropagation), unsupervised learning, reinforcement learning & evolutionary computation, and indirect search for short programs encoding deep and large networks.LATEX source: http://www.idsia.ch/˜juergen/DeepLearning8Oct2014.tex Complete BIBTEX file (888 kB): http://www.idsia.ch/˜juergen/deep.bib Preface This is the preprint of an invited Deep Learning (DL) overview. One of its goals is to assign credit to those who contributed to the present state of the art. I acknowledge the limitations of attempting to achieve this goal. The DL research community itself may be viewed as a continually evolving, deep network of scientists who have influenced each other in complex ways. Starting from recent DL results, I tried to trace back the origins of relevant ideas through the past half century and beyond, sometimes using "local search" to follow citations of citations backwards in time. Since not all DL publications properly acknowledge earlier relevant work, additional global search strategies were employed, aided by consulting numerous neural network experts. As a result, the present preprint mostly consists of references. Nevertheless, through an expert selection bias I may have missed important work. A related bias was surely introduced by my special familiarity with the work of my own DL research group in the past quarter-century. For these reasons, this work should be viewed as merely a snapshot of an ongoing credit assignment process. To help improve it, please do not hesitate to send corrections and suggestions to juergen@idsia.ch.

show abstract

Section: -1981 and Beyond: Development Of Backpropagation (Bp) Fomentioning

confidence: 99%

Deep learning in neural networks: An overview

2015

View full text Add to dashboard Cite

show abstract

“…The analogy with the sensitivity evaluation in electrical networks [5] suggested the use of an adjoint approach in dc and ac small-signal device sensitivity computation, as proposed in [6] within the framework of a majority-carrier drift-diffusion GaAs FET model. To allow for the extension of the method to bipolar and nonstationary transport models, some of the present-day authors proposed a generalization of the adjoint approach, whose idea is derived from Branin's method for the sensitivity analysis of electrical networks [7].…”

Section: Introductionmentioning

confidence: 99%

Physics-based large-signal sensitivity analysis of microwave circuits using technological parametric sensitivity from multidimensional semiconductor device models

Bonani

Guerrieri

Filicori

et al. 1997

IEEE Trans. Microwave Theory Techn.

View full text Add to dashboard Cite

Abstract-The authors present an efficient approach to evaluate the large-signal (LS) parametric sensitivity of active semiconductor devices under quasi-periodic operation through accurate, multidimensional physics-based models. The proposed technique exploits efficient intermediate mathematical models to perform the link between physics-based analysis and circuit-oriented simulations, and only requires the evaluation of dc and ac small-signal (dc charge) sensitivities under general quasi-static conditions. To illustrate the technique, the authors discuss examples of sensitivity evaluation, statistical analysis, and doping profile optimization of an implanted MESFET to minimize intermodulation which make use of LS parametric sensitivities under two-tone excitation.

show abstract

“…Another category is the sensitivity based optimization [15,1,9], in which the adjoint network method [5,4] can be applied to calculate sensitivity:…”

Section: Review Of Sensitivity-based Decap Allocation Algorithmsmentioning

confidence: 99%

“…A problem with this method is that decap area is not explicitly minimized but only taken as a constraint, which may lead to overestimated decaps. Moreover, sensitivity computation with the adjoint network method [5,4] is quite expensive in this case. Two simulations are required s ij for each node j, one for the normal circuit and one for the adjoint circuit with the same size.…”

Section: Review Of Sensitivity-based Decap Allocation Algorithmsmentioning

confidence: 99%

“…Two major methods to calculate incremental sensitivity are the direct method which is based on definition of sensitivity and involves construction of a sensitivity circuit, and the adjoint method [5,4], which is deducted from Tellegen's Theorem and involves construction of an adjoint circuit [12,7]. As concluded in [7], the former has advantage in computation of sensitivities of many responses with respect to a few parameters, and the latter fits better in cases where sensitivities of a few responses with respect to many parameters are required, which is exactly our case.…”

Section: Sensitivity Computationmentioning

confidence: 99%

See 1 more Smart Citation