on the Geometric Separability of Boolean Functions

Abstract We investigate the complexity of the MEMBERSHIP problem for some geometrically defined classes of Boolean functions, i.e., the complexity of deciding whether a Boolean function given in DNF belongs to the class. We give a general argument implying that this problem is co-NP-hard for any class having some rather benign closure properties. Applying this result we show that the MEMBERSHIP problem is co-NP-complete for the class of linearly separable functions, threshold functions of order k (for any fixed k ⩾ 0), and some binary-parameter analogues of these classes. Finally, we obtain that the considered problem for unions of k ⩾ 3 halfspaces is NP-hard, co-NP-hard and belongs to Σ 2 p , and that the optimal threshold decomposition of a Boolean function as a union of halfspaces cannot even be efficiently approximated in a very strong sense unless P = NP. In some cases we improve previous hardness results on the considered problems.

[1]  M. Yannakakis.  The Complexity of the Partial Order Dimension Problem , 1982 .

[2]  V. Klee,et al.  Helly's theorem and its relatives , 1963 .

[3]  Tibor Hegedüs.  Can Complexity Theory Benefit from Learning Theory? , 1993, ECML.

[4]  E. Helly.  Über Mengen konvexer Körper mit gemeinschaftlichen Punkte. , 1923 .

[5]  Yves Crama,et al.  Recognition problems for special classes of polynomials in 0–1 variables , 1989, Math. Program..

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

[7]  Lisa Hellerstein,et al.  Read-thrice DNF is hard to learn with membership and equivalence queries , 1992, Proceedings., 33rd Annual Symposium on Foundations of Computer Science.

[8]  György Turán,et al.  A Survey of Some Aspects of Computational Learning Theory (Extended Abstract) , 1991, FCT.

[9]  Toshihide Ibaraki,et al.  Threshold Numbers and Threshold Completions , 1981 .

[10]  Leonard Pitt,et al.  The minimum consistent DFA problem cannot be approximated within any polynomial , 1993, JACM.

[11]  Wolfgang Maass,et al.  How fast can a threshold gate learn , 1994, COLT 1994.

[12]  Leonard Pitt,et al.  The minimum consistent DFA problem cannot be approximated within any polynomial , 1989, [1989] Proceedings. Structure in Complexity Theory Fourth Annual Conference.

[13]  Tibor Hegedűs,et al.  On training simple neural networks and small-weight neurons , 1994 .

[14]  David S. Johnson,et al.  A Catalog of Complexity Classes , 1991, Handbook of Theoretical Computer Science, Volume A: Algorithms and Complexity.

[15]  P. Hammer,et al.  Aggregation of inequalities in integer programming. , 1975 .

[16]  J. Van Leeuwen,et al.  Handbook of theoretical computer science - Part A: Algorithms and complexity; Part B: Formal models and semantics , 1990 .

[17]  Leslie G. Valiant,et al.  Computational limitations on learning from examples , 1988, JACM.

[18]  Dana Angluin,et al.  Computational learning theory: survey and selected bibliography , 1992, STOC '92.

[19]  Chi Wang,et al.  The threshold order of a Boolean function , 1991, Discret. Appl. Math..

[20]  Bruno Simeone,et al.  Polynomial-time algorithms for regular set-covering and threshold synthesis , 1985, Discret. Appl. Math..

[21]  Wolfgang Maass,et al.  On the complexity of learning from counterexamples , 1989, 30th Annual Symposium on Foundations of Computer Science.

[22]  José L. Balcázar,et al.  Structural Complexity I , 1995, Texts in Theoretical Computer Science An EATCS Series.

[23]  M. Golumbic.  Algorithmic graph theory and perfect graphs , 1980 .

[24]  David S. Johnson,et al.  Computers and Intractability: A Guide to the Theory of NP-Completeness , 1978 .