Abstract
Boolean functions that have constant degree polynomial representation over a fixed finite ring form a natural and strict subclass of the complexity class ACC0. They are also precisely the functions computable efficiently by programs over fixed and finite nilpotent groups. This class is not known to be learnable in any reasonable learning model.
In this paper, we provide a deterministic polynomial time algorithm for learning Boolean functions represented by polynomials of constant degree over arbitrary finite rings from membership queries, with the additional constraint that each variable in the target polynomial appears in a constant number of monomials. Our algorithm extends to superconstant but low degree polynomials and still runs in quasipolynomial time.
Similar content being viewed by others
Notes
In an equivalence query, the learning algorithm emits some representation of a Boolean function and receives as answer either a Boolean assignment where it differs from the target function, or “yes” if no such assignment exists.
Alternatively one could fix the same ordering, say 1,…,n for all monomials. However we find it natural to identify monomials that are identical up to a permutation of the variables.
A DNF formula is read-k if every variable appears at most k times; a DNF formula is satisfy-j if no assignment satisfies more than j terms simultaneously.
References
Aizenstein, H., Blum, A., Khardon, R., Kushilevitz, E., Pitt, L., Roth, D.: On learning read-k-satisfy-j dnf. SIAM J. Comput. 27(6), 1515–1530 (1998)
Aizenstein, H., Hegedüs, T., Hellerstein, L., Pitt, L.: Complexity theoretic hardness results for query learning. Comput. Complex. 7(1), 19–53 (1998)
Angluin, D., Hellerstein, L., Karpinski, M.: Learning read-once formulas with queries. J. ACM 40(1), 185–210 (1993)
Barrington, D.A.M.: Bounded-width polynomial-size branching programs recognize exactly those languages in NC1. J. Comput. Syst. Sci. 38(1), 150–164 (1989)
Barrington, D.A.M., Beigel, R., Rudich, S.: Representing boolean functions as polynomials modulo composite numbers. Comput. Complex. 4, 367–382 (1994)
Barrington, D.A.M., Straubing, H.: Superlinear lower bounds for bounded-width branching programs. J. Comput. Syst. Sci. 50(3), 374–381 (1995)
Barrington, D.A.M., Straubing, H., Thérien, D.: Non-uniform automata over groups. Inf. Comput. 89(2), 109–132 (1990)
Barrington, D.A.M., Thérien, D.: Finite monoids and the finite structure of NC1. J. ACM 35(4), 941–952 (1988)
Beimel, A., Bergadano, F., Bshouty, N., Kushilevitz, E., Varricchio, S.: Learning functions represented as multiplicity automata. J. ACM 47, 506–530 (2000)
Bourgain, J.: Estimation of certain exponential sums arising in complexity theory. C. R. Math. Acad. Sci. 340(9), 627–631 (2005)
Braverman, M., Rao, A., Raz, R., Yehudayoff, A.: Pseudorandom generators for regular branching programs. In: 51th Annual IEEE Symposium on Foundations of Computer Science, pp. 40–47. IEEE Comput. Soc., Los Alamitos (2010)
Brody, J., Verbin, E.: The coin problem, and pseudorandomness for branching programs. In: 51th Annual IEEE Symposium on Foundations of Computer Science, pp. 30–39. IEEE Comput. Soc., Los Alamitos (2010)
Bshouty, N., Tamon, C., Wilson, D.: Learning matrix functions over rings. Algorithmica 22, 91–111 (1998)
Chattopadhyay, A.: Multilinear polynomials modulo composites. In: Bulletin of the European Association on Theoretical Computer Science, Computational Complexity Column, vol. 100 (2010)
Efremenko, K.: 3-query locally decodable codes of subexponential length. In: 41st Annual Symposium on Theory of Computing, pp. 39–44. ACM, New York (2009)
Gavaldà, R., Thérien, D.: An algebraic perspective on boolean function learning. In: Algorithmic Learning Theory, 20th International Conference, ALT 2009. LNCS, pp. 201–215. Springer, Berlin (2009)
Gopalan, P.: Constructing Ramsey graphs from boolean function representations. In: 21st Annual IEEE Conference on Computational Complexity, pp. 115–128. IEEE Comput. Soc., Los Alamitos (2006)
Grolmusz, V.: On the weak mod m representation of boolean functions. Chic. J. Theor. Comput. Sci. (1995)
Helmbold, D.P., Sloan, R.H., Warmuth, M.K.: Learning nested differences of intersection-closed concept classes. Mach. Learn. 5, 165–196 (1990)
Koucký, M., Nimbhorkar, P., Pudlák, P.: Pseudorandom generators for group products. In: 43rd Annual ACM Symposium on Theory of Computing, pp. 263–272. ACM, New York (2011)
Péladeau, P., Thérien, D.: Sur les langages reconnus par des groupes nilpotents. C. R. Math. Acad. Sci. Paris 306(2), 93–95 (1988)
Péladeau, P., Thérien, D.: On the languages recognized by nilpotent groups (a translation of “Sur les langages reconnus par des groupes nilpotents” [21]). Electron. Colloq. Comput. Complex. 8(40) (2001)
Pillaipakkamnatt, K., Raghavan, V.V.: Read-twice dnf formulas are properly learnable. Inf. Comput. 122(2), 236–267 (1995)
Razborov, A.A.: Lower bounds on the size of bounded-depth networks over a complete basis with logical addition. Math. Notes Acad. Sci. USSR 41(3), 333–338 (1987)
Schapire, R.E., Sellie, L.: Learning sparse multivariate polynomials over a field with queries and counterexamples. J. Comput. Syst. Sci. 52(2), 201–213 (1996)
Smolensky, R.: Algebraic methods in the theory of lower bounds for boolean circuit complexity. In: 19th Annual ACM Symposium on Theory of Computing, pp. 77–82. ACM, New York (1987)
Tardos, G., Barrington, D.A.M.: A lower bound on the MOD-6 degree of the OR function. Comput. Complex. 7(2), 99–108 (1998)
Acknowledgements
A. Chattopadhyay was at the University of Toronto when this work was written up, partially supported by a Natural Sciences and Engineering Research Council (NSERC) postdoctoral fellowship, an Ontario Ministry of Research and Innovation fellowship and research grants of Prof. T. Pitassi. R. Gavaldà is partially funded by the Spanish MICINN projects TIN-2008-06582-C03-01 (SESAAME) and TIN2011-27479-C04-03 (BASMATI), by the Generalitat de Catalunya SGR2009-1428 (LARCA), and by the EU PASCAL2 Network of Excellence (FP7-ICT-216886). We thank the anonymous reviewers for their useful comments that improved the presentation. In particular, Remark 5 was pointed out by a reviewer.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Chattopadhyay, A., Gavaldà, R., Hansen, K.A. et al. Learning Read-Constant Polynomials of Constant Degree Modulo Composites. Theory Comput Syst 55, 404–420 (2014). https://doi.org/10.1007/s00224-013-9488-6
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00224-013-9488-6