Skip to main content
Log in

A divisibility approach to the open boundary cases of Cusick-Li-Stǎnicǎ’s conjecture

  • Published:
Cryptography and Communications Aims and scope Submit manuscript

Abstract

In this paper we compute the exact 2-divisibility of exponential sums associated to elementary symmetric Boolean functions. Our computation gives an affirmative answer to most of the open boundary cases of Cusick-Li-Stǎnicǎ’s conjecture. As a byproduct, we prove that the 2-divisibility of these families satisfies a linear recurrence. In particular, we provide a new elementary method to compute 2-divisibility of symmetric Boolean functions.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8

Similar content being viewed by others

References

  1. Adolphson, A., Sperber, S.: p-adic estimates for exponential sums and the of chevalley-warning. Ann. Sci. Ec. Norm. Super., 4e série 20, 545–556 (1987)

    MathSciNet  MATH  Google Scholar 

  2. Ax, J.: Zeros of polynomials over finite fields. Amer. J. Math. 86, 255–261 (1964)

    Article  MathSciNet  MATH  Google Scholar 

  3. Bassalygo, L.A., Zinoviev, V.A.: On divisibility of exponential sums of polynomials of special type over fields of characteristic 2. Des. Codes Cryptogr 66, 129–143 (2013)

    Article  MathSciNet  MATH  Google Scholar 

  4. Canteaut, A., Charpin, P., Dobbertin, H.: Weight divisibility of cyclic codes, highly nonlinear functions on G F(2m) and cross correlation of maximum-length sequences. SIAM J. Discret. Math. 13, 105–138 (2000)

    Article  MathSciNet  MATH  Google Scholar 

  5. Castro, F., Medina, L.: Linear recurrences and asymptotic behavior of exponential sums of symmetric boolean functions. Elec. J. Combinatorics 18, #P8 (2011)

    MathSciNet  Google Scholar 

  6. Castro, F., Medina, L.: Asymptotic behavior of perturbations of symmetric functions. Ann. Comb. 18, 397–417 (2014)

    Article  MathSciNet  MATH  Google Scholar 

  7. Castro, F., Medina, L., Rubio, I.: Exact Divisibility of exponential sums over the binary field via the covering method. Contemp. Math. 537, 129–136 (2009)

    Article  MathSciNet  Google Scholar 

  8. Castro, F., Rubio, I.: Exact p-divisibility of exponential sums via the covering method. Proc. Amer. Math. Soc., electronically published on October 29,2014, doi: 10.1090/S0002-9939-2014-12315-X (to appear in print)

  9. Castro, F., Rubio, I.: Construction of Systems of polynomial equations with exact p-divisibility via the covering method. J. Algebra Appl. 13 (6), 1450013 (2014)

    Article  MathSciNet  Google Scholar 

  10. Cusick, T.W., Li, Y., Stǎnicǎ, P. IEEE Trans. Inf. Theory 5, 1304–1307 (2008)

    Article  MathSciNet  Google Scholar 

  11. Cusick, T.W., Li, Y., Stǎnicǎ, P.: On a conjecture for balanced symmetric Boolean functions. J. Math. Crypt. 3, 1–18 (2009)

    Article  MathSciNet  Google Scholar 

  12. Davis, K.S., Webb, W.A.: Lucas’ theorem for prime powers. Europ. J. Comb. 11, 229–233 (1990)

    Article  MathSciNet  MATH  Google Scholar 

  13. Guo, Y., Gao, G., Zhao, Y.: Recent Results on Balanced Symmetric Boolean Functions. http://iacr.org, e-print #93 (2012)

  14. Granville, A.: Zaphod Beeblerox’s brain and the fifty-ninth row of Pascal’s triangle. American. Math. Monthly 44, 318–331 (1992)

    Article  MathSciNet  Google Scholar 

  15. Güneri, C., McGuire, G.: Supersingular curves over finite fields and weight divisibility of codes. J. Comput. Appl. Math. 259, part B, 474–484 (2014)

    Article  Google Scholar 

  16. Huard, J.G., Spearman, B.K., Williams, K.: Pascal triangle (mod 8). Europ. J. Comb. 19, 45–61 (1998)

    Article  MathSciNet  MATH  Google Scholar 

  17. Gao, G-P., Liu, W-F, Zhang, X-Y.: The degree of balanced elementary symmetric Boolean functions of 4k+3 variables. IEEE Trans. Inf. Theory 57, 4822–4825 (2011)

    Article  MathSciNet  Google Scholar 

  18. Kolountzakis, M., Lipton, R.J., Markakis, E., Metha, A., Vishnoi, N.K.: On the fourier spectrum of symmetric Boolean functions. Combinatorica 29, 363–387 (2009)

    Article  MathSciNet  MATH  Google Scholar 

  19. Moreno, O., Castro, F.: Divisibility properties for covering radius of certain cyclic codes. IEEE Trans. Inform. Theory 49 (12), 3299–3303 (2003)

    Article  MathSciNet  Google Scholar 

  20. Moreno, O., Moreno, C.J.: Improvement of the Chevalley-warning and the Ax-Katz theorems. Amer. J. Math 117, 241–244 (1995)

    Article  MathSciNet  MATH  Google Scholar 

  21. Moreno, O., Moreno, C.J.: The MacWilliams-Sloane conjecture on the tightness of the Carlitz-Uchiyama bound and the weights of dual of BCH codes. IEEE Trans. Inform. Theory 40, 1894–1907 (1994)

    Article  MathSciNet  MATH  Google Scholar 

  22. Moreno, O., Shum, K., Castro, F.N., Kumar, P.V.: Tight bounds for Chevalley-Warning-Ax type estimates, with improved applications. Proc. London Math. Soc. 88, 545–564 (2004)

    Article  MathSciNet  MATH  Google Scholar 

  23. Shpilka, A., Tal, A.: On the minimal fourier degree of symmetric Boolean functions. Combinatorica 88, 359–377 (2014)

    Article  MathSciNet  Google Scholar 

  24. von zur Gathem, J., Roche, J.R.: Polynomial with two values. Combinatorica 17, 345–362 (1997)

    Article  MathSciNet  Google Scholar 

  25. Wei, S., Tang, X., Pott, A.: A note on a conjecture for balanced elementary symmetric Boolean functions. IEEE Trans. Inf. Theory 59, 665–671 (2013)

    Article  Google Scholar 

Download references

Acknowledgments

We would like to thank Professor Thomas W. Cusick for his helpful comments and suggestions in a previous version of this paper and to the referees for improving the presentation of this manuscript. The second author was partially supported as a student by NSF-DUE 1356474. The third author acknowledges the partial support of UPR-FIPI 1890015.00.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Luis A. Medina.

Additional information

The second author was partially supported as a student by NSF-DUE 1356474. The third author acknowledges the partial support of UPR-FIPI 1890015.00.

Appendix A: Proof of Lemma 2

Appendix A: Proof of Lemma 2

This appendix is devoted to the proof of Lemma 2. We re-state the lemma in order to aid the reader.

Lemma 2

Let a,i,m be natural numbers with i≥3 and m odd. Write m=b s ⋅2 s +b s−1 ⋅2 s−1 +⋯+b 1 ⋅2+1 with s≥1. Let \(b_{s - l_{1}}, b_{s - l_{2}}, \ldots , b_{s - l_{r}}\) be all the b t in the expansion of m such that b t =0. Define

$$g_{a,i}=\delta_{1} \cdot 2^{a+i-1+(s - l_{1})} + \cdots + \delta_{r} \cdot 2^{a+i-1+(s - l_{r})} $$

Then,

$$ \frac{(2^{a+i-1} \cdot m - 2^{a} +g_{a,i} + 2^{a+i-1}) \cdots (2^{a+i-1} \cdot m - 2^{a} +g_{a,i}+1)} {(2^{a+i-1} \cdot m - g_{a,i}) \cdots (2^{a+i-1} \cdot m - g_{a,i} -2^{a+i-1} + 1)} \equiv 3 \;\text{mod}\;{4}. $$
(70)

Proof

The proof we present is elementary, but rather long. We decided to present most of the details, including a double induction, because this technique can be used to prove other results.

Note that the left hand side of (70) is an odd number, which implies that the numbers of even terms in the numerator and the denominator are equal. This will allow us to perform reductions modulo 8 in the numerator and denominator independently. Also note that the number of terms in the numerator and denominator is 2a+i−1. We prove (70) by double induction.

Suppose first that i=3. We want to prove that

$$\frac{(2^{a+2} \cdot m - 2^{a} +g_{a,3} + 2^{a+2}) \cdots (2^{a+2} \cdot m - 2^{a} +g_{a,3}+1)} {(2^{a+2} \cdot m - g_{a,3}) \cdots (2^{a+2} \cdot m - g_{a,3} -2^{a+2} + 1)} \equiv 3 \;\text{mod}\;{4} $$

for a≥1. For the base case, i.e. a=1, we have

$$\begin{array}{@{}rcl@{}} && \frac{(2^{3} \cdot m - 2 +g_{1,3} + 2^{3}) \cdots (2^{3} \cdot m - 2 +g_{1,3}+1)} {(2^{3} \cdot m - g_{1,3}) \cdots (2^{3} \cdot m - g_{1,3} -2^{3} + 1)}\\ &&= \frac{(2^{3} \cdot m +g_{1,3} + 6) \cdots (2^{3} \cdot m +g_{1,3} -1)} {(2^{3} \cdot m - g_{1,3}) \cdots (2^{3} \cdot m - g_{1,3} -7)}\\ &&= \frac{(2^{3} \cdot m +g_{1,3} + 6) \cdots (2^{3} \cdot m +g_{1,3} +1) \cdot (2^{3} \cdot m +g_{1,3} -1)} {(2^{3} \cdot m - g_{1,3} -1) \cdots (2^{3} \cdot m - g_{1,3} -7)} \cdot \frac{(2^{3} \cdot m +g_{1,3})}{(2^{3} \cdot m - b)}\\ &&\equiv \frac{(4 \cdot m + 3) \cdot (2 \cdot m +1) \cdot (4 \cdot m +1)} {(4 \cdot m -1) \cdot (2 \cdot m -1) \cdot (4 \cdot m - 3)} \\ && \,\,\,\,\,\,\, \times \frac{(m + \delta_{1} \cdot 2^{(s - l_{1})} + \cdots + \delta_{r} \cdot 2^{(s - l_{r})})} {(m - \delta_{1} \cdot 2^{(s - l_{1})} - \cdots - \delta_{r} \cdot 2^{(s - l_{r})})} \;\text{mod}\;{8}\\ &&\equiv 3 \;\text{mod}\;{4}. \end{array} $$

Thus the claim holds for a=1. Suppose that

$$ \frac{(2^{a+2} \cdot m - 2^{a} +g_{a,3} + 2^{a+2} ) \cdots (2^{a+2} \cdot m - 2^{a} +g_{a,3}+1)} {(2^{a+2} \cdot m - g_{a,3}) \cdots (2^{a+2} \cdot m - g_{a,3} -2^{a+2} + 1)} \equiv 3 \;\text{mod}\;{4} $$
(71)

for some a≥1. We will show that

$$ \frac{(2^{a+3} \cdot m - 2^{a+1} +g_{a+1,3} + 2^{a+3}) \cdots (2^{a+3} \cdot m - 2^{a+1} +g_{a+1,3}+1)} {(2^{a+3} \cdot m - g_{a+1,3}) \cdots (2^{a+3} \cdot m - g_{a+1,3} -2^{a+3} + 1)} \equiv 3 \;\text{mod}\;{4}. $$
(72)

Observe that the numerator and denominator of the left hand side of (72) have 2a+2 consecutive odd terms. Thus, the left hand side of (72) is congruent modulo 4 to

$$\begin{array}{@{}rcl@{}} \frac{2^{a+3} \cdot m - 2^{a+1} +g_{a+1,3} + 2^{a+3}}{2^{a+3} \cdot m - g_{a+1,3}} \times \frac{2^{a+3} \cdot m - 2^{a+1} +g_{a+1,3} + 2^{a+3}-2}{2^{a+3} \cdot m - g_{a+1,3}-2} \\ \times \cdots \times \frac{2^{a+3} \cdot m - 2^{a+1} +g_{a+1,3}}{2^{a+3} \cdot m - g_{a+1,3} -2^{a+3}} \end{array} $$

Now factor a 2 out of each term to obtain

$$ \frac{(2^{a+2} \cdot m - 2^{a} +g_{a+1,3}/2 + 2^{a+2}) \cdots (2^{a+2} \cdot m - 2^{a} +g_{a+1,3}/2)} {(2^{a+2} \cdot m - g_{a+1,3}/2) \cdots (2^{a+2} \cdot m - g_{a+1,3}/2 -2^{a+2})} . $$
(73)

Since g a+1,3/2=g a,3, then (73) is the same number as the left hand side of (71) and so, by our induction hypothesis, (72) is congruent to 3 mod 4, which is what we wanted to prove. This takes care of the first step of the double induction.

Suppose that

$$ \frac{(2^{a+i-1} \cdot m - 2^{a} +g_{a,i} + 2^{a+i-1}) \cdots (2^{a+i-1} \cdot m - 2^{a} +g_{a,i}+1)} {(2^{a+i-1} \cdot m - g_{a,i}) \cdots (2^{a+i-1} \cdot m - g_{a,i} -2^{a+i-1} + 1)} \equiv 3 \;\text{mod}\;{4}. $$
(74)

is true for some i≥3. Consider the case i+1, i.e.

$$\frac{(2^{a+i} \cdot m - 2^{a} +g_{a,i+1} + 2^{a+i}) \cdots (2^{a+i} \cdot m - 2^{a} +g_{a,i+1}+1)} {(2^{a+i} \cdot m - g_{a,i+1}) \cdots (2^{a+i} \cdot m - g_{a,i+1} -2^{a+i} + 1)}. $$

Note that there are 2a+i consecutive integers in the numerator and in the denominator. Thus, we have 2a+i−1 consecutive odd numbers in the numerator and in the denominator. Since a+i−1≥2, then we have that these terms will be congruent to 1 mod 8. After this cancellation, we are left with the even terms:

$$ \frac{(2^{a+i} \cdot m - 2^{a} +g_{a,i+1} + 2^{a+i}) \cdots (2^{a+i} \cdot m - 2^{a} +g_{a,i+1}+2)} {(2^{a+i} \cdot m - g_{a,i+1}) \cdots (2^{a+i} \cdot m - g_{a,i+1} -2^{a+i}+2)}. $$
(75)

Observe that if a≥2, then, after taking this expression mod 4, we have

$$ \frac{(2^{a+i} \cdot m +g_{a,i+1} + 2^{a+i}) \cdots (2^{a+i} \cdot m +g_{a,i+1}+2)} {(2^{a+i} \cdot m - g_{a,i+1}) \cdots (2^{a+i} \cdot m - g_{a,i+1} -2^{a+i}+2)} \;\text{mod}\;{4}. $$
(76)

Now factor a 2 out of each term to obtain

$$ \frac{(2^{a+i-1} \cdot m +g_{a,i} + 2^{a+i-1}) \cdots (2^{a+i-1} \cdot m +g_{a,i}+1)} {(2^{a+i-1} \cdot m - g_{a,i}) \cdots (2^{a+i-1} \cdot m - g_{a,i} -2^{a+i-1}+1)}\;\text{mod}\;{4}, $$
(77)

and this expression is congruent to 3 modulo 4 by induction. If a=1, then observe that (74) gets transformed to

$$ \frac{(2^{i} \cdot m +g_{1,i} + 2^{i}-2)(2^{i} \cdot m +g_{1,i} + 2^{i}-1) \cdots (2^{i} \cdot m +g_{1,i}-1)} {(2^{i} \cdot m - g_{1,i}) \cdots (2^{i} \cdot m - g_{1,i} -2^{i} + 1)} \equiv 3 \;\text{mod}\;{4}, $$
(78)

while (75) gets transformed to

$$\frac{(2^{i+1} \cdot m +g_{1,i+1} + 2^{i+1}-2) \cdots (2^{i+1} \cdot m +g_{1,i+1})} {(2^{i+1} \cdot m - g_{1,i+1}) \cdots (2^{i+1} \cdot m - g_{1,i+1} -2^{i+1}+2)}. $$
(79)

Factor a 2 out of each term to obtain

$$\frac{(2^{i} \cdot m +g_{1,i} + 2^{i}-1) \cdots (2^{i} \cdot m +g_{1,i})} {(2^{i} \cdot m - g_{1,i}) \cdots (2^{i} \cdot m - g_{1,i} -2^{i}+1)}. $$
(80)

Since 2i−1≡−1 mod 8, then (80) is congruent to (78) modulo 4, and thus, congruent to 3 mod 4 by induction. This concludes the proof.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Castro, F.N., González, O.E. & Medina, L.A. A divisibility approach to the open boundary cases of Cusick-Li-Stǎnicǎ’s conjecture. Cryptogr. Commun. 7, 379–402 (2015). https://doi.org/10.1007/s12095-015-0124-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s12095-015-0124-y

Keywords

Mathematical Subject Classifications (2010)

Navigation