On the inverse Legendre transform of a certain family of sequences (Q1889712)

Here the author proves the following theorem: For an integer \(r\geq2,\) let the sequence of numbers \(\{c_ {k}^ {(r)}(n)\}_ {k=0}^ \infty\), depending on the parameter \(n\), be given by the equation \[ \sum_ {k=0}^ n\left({n\atop k}\right)^ r\left({n+k\atop k}\right)^ r= \sum_ {k=0}^ n\left({n\atop k}\right)\left({n+k\atop k}\right)c_ k^ {(r)}(n),\;n=0,1,\dots \,. \] Then all the numbers \(c_ k^ {(r)}(n)\) are integers; they are calculated explicitly. It is pointed out that the sequence \(\{c_ k^ {(r)}\}\) in the case \(r=2\) gives the inverse Legendre transform of the famous sequence of Apéry's numbers expressing the denominators of the convergents in his proof of the irrationality of the number \(\zeta(3)=\sum_{n=1}^\infty n^{-3}\). The problem, originally posed by \textit{A. L. Schmidt} [J. Comput. Appl. Math. 49, No.1-3, 243--249 (1993; Zbl 0797.33013)] was solved by Schmidt for \(r=2\) and by \textit{V. Strehl} [Discrete Math. 136, No. 1-3, 309--346 (1994; Zbl 0823.33003)] for \(r=2,3\).