The slicing number of a knot (Q1855203)

Let \(K\) be a knot in the \(3\)-sphere \({\mathbf S}^3=\partial{\mathbf B}^4\). The \textsl{\(4\)-genus} of \(K\) is the minimal genus of all orientable embedded surfaces bounded by \(K\) in \({\mathbf B}^4\). \(K\) is \textsl{slice} if its \(4\)-genus is \(0\). The \textsl{slicing number} of \(K\) is the minimal number of crossing changes necessary to transform \(K\) into a slice knot. It is easy to see that the slicing number of a knot is bounded below by its \(4\)-genus. The paper answers in the negative the following question of Askitas: do the slicing number and the \(4\)-genus of a knot always coincide? (as it is the case for slice knots). Indeed, the author shows that the knot \(7_4\) has slicing number \(2\) but \(4\)-genus \(1\). In this case, the slicing number coincides with the unkotting number [this fact was proved by \textit{R. Lickorish} in Contemp. Math. 44, 117-121 (1985; Zbl 0607.57002)], and the \(4\)-genus with the usual one. To prove the result, the author uses the classical fact that a crossing change can be obtained by performing a surgery. He then analyses the double and infinite cyclic covers of the knot obtained after some crossing change. These covers can be recovered from those of the original knot via surgeries. To conclude, the author introduces a new slicing invariant that takes into account the type of crossing changes that are performed and observes that this invariant is also bounded below by the \(4\)-genus of a knot. The following questions are then natural: 1) Can the difference between the slicing number and the \(4\)-genus be arbitrarily large? 2) What can be said of the difference between the new slicing invariant and the \(4\)-genus?