Title: Formal Power Series on Algebraic Cryptanalysis

Abstract: In cryptography, by attacks reducing a cryptosystem to the problem that solves the system of polynomial equations, several cryptosystems have been broken. The efficiency of an algorithm to solve the polynomial system depends on the solving degree, and its complexity is estimated using a theoretical proxy for the solving degree. If a polynomial system generated by an attack is semi-regular, the degree of regularity determined from a certain univariate formal power series is used as such a proxy, otherwise, the first fall degree is used. In this article, we mainly investigate an upper bound for the proxy defined using non-constructive syzygies such as the first fall degree. In particular, for a sufficiently large field, we prove that the first fall degree of a non-semi-regular system is bounded above by the degree of regularity of the semi-regular system, and that the first fall degree of a multi-graded polynomial system is bounded above by a certain value determined from a multivariate formal power series. Moreover, we show that the assumption for the order of a coefficient field in our results is satisfied in actual attacks against multivariate cryptosystems. Furthermore, we provide the theoretical assumption for the algorithm with a kernel search to solve a multi-graded polynomial system and compute a certain proxy for the solving degree by the multivariate formal power series. Consequently, we clarify the relationship between the first fall degree and the degree of regularity over a sufficiently large field, and provide a theoretical method using a multivariate power series for cryptanalysis.