A New Decomposition Algorithm for Threshold Synthesis and Generalization of Boolean Functions

A new algorithm for obtaining efficient architectures composed of threshold gates that implement arbitrary Boolean functions is introduced. The method reduces the complexity of a given target function by splitting the function according to the variable with the highest influence. The procedure is iteratively applied until a set of threshold functions is obtained, leading to reduced depth architectures, in which the obtained threshold functions form the nodes and a and or or function is the output of the architecture. The algorithm is tested on a large set of benchmark functions and the results compared to previous existing solutions, showing a considerable reduction on the number of gates and levels of the obtained architectures. An extension of the method for partially defined functions is also introduced and the generalization ability of the method is analyzed.

[1]  Alberto L. Sangiovanni-Vincentelli,et al.  LSAT-an algorithm for the synthesis of two level threshold gate networks , 1991, 1991 IEEE International Conference on Computer-Aided Design Digest of Technical Papers.

[2]  Aiko M. Hormann,et al.  Programs for Machine Learning. Part I , 1962, Inf. Control..

[3]  Leonardo Franco,et al.  Generalization ability of Boolean functions implemented in feedforward neural networks , 2006, Neurocomputing.

[4]  Ingo Wegener.  The Complexity of the Parity Function in Unbounded Fan-In, Unbounded Depth Circuits , 1991, Theor. Comput. Sci..

[5]  Martin Anthony,et al.  The influence of oppositely classified examples on the generalization complexity of Boolean functions , 2006, IEEE Transactions on Neural Networks.

[6]  Anthony N. Michel,et al.  A training algorithm for binary feedforward neural networks , 1992, IEEE Trans. Neural Networks.

[7]  Leonardo Franco,et al.  Neural Network Architecture Selection: Size Depends on Function Complexity , 2006, ICANN.

[8]  P. R. Stephan,et al.  SIS : A System for Sequential Circuit Synthesis , 1992 .

[9]  J. Nadal,et al.  Learning in feedforward layered networks: the tiling algorithm , 1989 .

[10]  Sze Tsen Hu,et al.  Threshold Logic , 1966 .

[11]  John E. Hopcroft,et al.  Synthesis of Minimal Threshold Logic Networks , 1965, IEEE Trans. Electron. Comput..

[12]  Eduardo Sontag,et al.  A Comparison of the Computational Power of Sigmoid and Boolean Threshold Circuits , 1994 .

[13]  G. Barkema,et al.  A Fast Partitioning Algorithm and a Comparison of Binary Feedforward Neural Networks , 1992 .

[14]  Sze-Tsen Hu.  ON THE DECOMPOSITION OF SWITCHING FUNCTIONS , 1961 .

[15]  Valeriu Beiu,et al.  VLSI implementations of threshold logic-a comprehensive survey , 2003, IEEE Trans. Neural Networks.

[16]  H. A. Curtis,et al.  A new approach to The design of switching circuits , 1962 .

[17]  Martin Anthony,et al.  A New Constructive Approach for Creating All Linearly Separable (Threshold) Functions , 2006, The 2006 IEEE International Joint Conference on Neural Network Proceedings.

[18]  Saburo Muroga,et al.  The principle of majority decision logical elements and the complexity of their circuits , 1959, IFIP Congress.

[19]  Sarma B. K. Vrudhula,et al.  Combinational equivalence checking for threshold logic circuits , 2007, GLSVLSI '07.

[20]  Kai-Yeung Siu,et al.  On Optimal Depth Threshold Circuits for Multiplication and Related Problems , 1994, SIAM J. Discret. Math..

[21]  Thomas Kailath,et al.  Depth-Size Tradeoffs for Neural Computation , 1991, IEEE Trans. Computers.

[22]  Marcus Frean,et al.  The Upstart Algorithm: A Method for Constructing and Training Feedforward Neural Networks , 1990, Neural Computation.

[23]  Hubertus M. A. Andree,et al.  A comparison study of binary feedforward neural networks and digital circuits , 1993, Neural Networks.

[24]  Michael L. Dertouzos,et al.  Threshold Logic: A Synthesis Approach , 1965 .

[25]  D. Liberati,et al.  Training digital circuits with Hamming clustering , 2000 .

[26]  Radomir S. Stankovic,et al.  DEDICATED SPECTRAL METHOD OF BOOLEAN FUNCTION DECOMPOSITION , 2006 .

[27]  Toshihide Ibaraki,et al.  An Implementation of Logical Analysis of Data , 2000, IEEE Trans. Knowl. Data Eng..

[28]  Valeriu Beiu,et al.  On the Circuit Complexity of Sigmoid Feedforward Neural Networks , 1996, Neural Networks.

[29]  Ingo Wegener,et al.  The complexity of Boolean functions , 1987 .

[30]  Robert O. Winder,et al.  Threshold logic , 1971, IEEE Spectrum.

[31]  Leonardo Franco,et al.  Generalization properties of modular networks: implementing the parity function , 2001, IEEE Trans. Neural Networks.

[32]  José M. Quintana,et al.  A threshold logic synthesis tool for RTD circuits , 2004 .

[33]  Saburo Muroga,et al.  Threshold logic and its applications , 1971 .

[34]  Se June Hong.  R-MINI: An Iterative Approach for Generating Minimal Rules from Examples , 1997, IEEE Trans. Knowl. Data Eng..

[35]  Michael E. Saks,et al.  Size-depth trade-offs for threshold circuits , 1993, SIAM J. Comput..

[36]  Niraj K. Jha,et al.  Threshold network synthesis and optimization and its application to nanotechnologies , 2005 .

[37]  Ian Parberry,et al.  Circuit complexity and neural networks , 1994 .

[38]  Paul Hasler,et al.  Programmable neural logic , 1998, 1997 Proceedings Second Annual IEEE International Conference on Innovative Systems in Silicon.

[39]  Eddy Mayoraz,et al.  Constructive Training Methods for feedforward Neural Networks with Binary weights , 1995, Int. J. Neural Syst..

[40]  Reiner Kolla,et al.  TROY: a tree based approach to logic synthesis and technology mapping , 1996, Proceedings of the Sixth Great Lakes Symposium on VLSI.

[41]  Leonardo Franco,et al.  Optimal Synthesis of Boolean Functions by Threshold Functions , 2006, ICANN.

[42]  Chee Kheong Siew,et al.  Can threshold networks be trained directly? , 2006, IEEE Transactions on Circuits and Systems II: Express Briefs.