Constant depth circuits, Fourier transform, and learnability

In this paper, Boolean functions in ,4C0 are studied using harmonic analysis on the cube. The main result is that an ACO Boolean function has almost all of its "power spectrum" on the low-order coefficients. An important ingredient of the proof is Hastad's switching lemma (8). This result implies several new properties of functions in -4C(': Functions in AC() have low "average sensitivity;" they may be approximated well by a real polynomial of low degree and they cannot be pseudorandom function generators. Perhaps the most interesting application is an O(n POIYIOg(n ')-time algorithm for learning func- tions in ACO. The algorithm observes the behavior of an AC'" function on O(nPO'Y'Og(n)) randomly chosen inputs, and derives a good approximation for the Fourier transform of the function. This approximation allows the algorithm to predict, with high probability, the value of the function on other randomly chosen inputs.

[1]  Harry Dym,et al.  Fourier series and integrals , 1972 .

[2]  Michael Sipser,et al.  Parity, circuits, and the polynomial-time hierarchy , 1981, 22nd Annual Symposium on Foundations of Computer Science (sfcs 1981).

[3]  Andrew Chi-Chih Yao,et al.  Theory and application of trapdoor functions , 1982, 23rd Annual Symposium on Foundations of Computer Science (sfcs 1982).

[4]  Miklós Ajtai,et al.  ∑11-Formulae on finite structures , 1983, Ann. Pure Appl. Log..

[5]  Leslie G. Valiant,et al.  A theory of the learnable , 1984, STOC '84.

[6]  Andrew Chi-Chih Yao,et al.  Separating the Polynomial-Time Hierarchy by Oracles (Preliminary Version) , 1985, FOCS.

[7]  Avi Wigderson,et al.  Deterministic simulation of probabilistic constant depth circuits , 1985, 26th Annual Symposium on Foundations of Computer Science (sfcs 1985).

[8]  A. Yao.  Separating the polynomial-time hierarchy by oracles , 1985 .

[9]  Silvio Micali,et al.  How to construct random functions , 1986, JACM.

[10]  J. Håstad.  Computational limitations of small-depth circuits , 1987 .

[11]  A. Razborov.  Lower bounds on the size of bounded depth circuits over a complete basis with logical addition , 1987 .

[12]  Roman Smolensky,et al.  Algebraic methods in the theory of lower bounds for Boolean circuit complexity , 1987, STOC.

[13]  Nathan Linial,et al.  The influence of variables on Boolean functions , 1988, [Proceedings 1988] 29th Annual Symposium on Foundations of Computer Science.

[14]  M. Kearns,et al.  Crytographic limitations on learning Boolean formulae and finite automata , 1989, STOC '89.

[15]  Torben Hagerup,et al.  A Guided Tour of Chernoff Bounds , 1990, Inf. Process. Lett..

[16]  Alon Orlitsky,et al.  A Spectral Lower Bound Techniqye for the Size of Decision Trees and Two Level AND/OR Circuits , 1990, IEEE Trans. Computers.

[17]  Noam Nisan,et al.  The computational complexity of universal hashing , 1990, STOC '90.

[18]  Noam Nisan,et al.  Approximate Inclusion-Exclusion , 1990, STOC '90.

[19]  Miklos Santha,et al.  Polynomial Size Constant Depth Circuits with a Limited Number of Negations , 1991, STACS.

[20]  Leslie G. Valiant,et al.  Cryptographic Limitations on Learning Boolean Formulae and Finite Automata , 1993, Machine Learning: From Theory to Applications.

[21]  Noam Nisan,et al.  The computational complexity of universal hashing , 1990, Proceedings Fifth Annual Structure in Complexity Theory Conference.

[22]  Noam Nisan,et al.  Hardness vs Randomness , 1994, J. Comput. Syst. Sci..