Georgia Institute of Technology College of Sciences

Knotted trivalent graphs (KTGs) along with standard operations defined on them form a finitely presented algebraic structure which includes knots, and in which many topological knot properties are defineable using simple formulas. Thus, a homomorphic invariant of KTGs places knot theory in an algebraic context. In this talk we construct such an invariant: the starting point is extending the Kontsevich integral of knots to KTGs. This was first done in a series of papers by Le, Murakami, Murakami and Ohtsuki in the late 90's using the theory of associators. We present an elementary construction building on Kontsevich's original definition, and discuss the homomorphic properties of the invariant, which, as it turns out, intertwines all the standard KTG operations except for one, called the edge unzip. We prove that in fact no universal finite type invariant of KTGs can intertwine all the standard operations at once, and present an alternative construction of the space of KTGs on which a homomorphic universal finite type invariant exists. This space retains all the good properties of the original KTGs: it is finitely presented, includes knots, and is closely related to Drinfel'd associators. (Partly joint work with Dror Bar-Natan.)

This talk is about the dilatations of pseudo-Anosov mapping classes obtained by pushing a marked point around a filling curve. After reviewing this "point-pushing" construction, I will give both upper and lower bounds on the dilatation in terms of the self-intersection number of the filling curve. I'll also give bounds on the least dilatation of any pseudo-Anosov in the point-pushing subgroup and describe the asymptotic dependence on self-intersection number. All of the upper bounds involve analyzing explicit examples using train tracks, and the lower bound is obtained by lifting to the universal cover and studying the images of simple closed curves.

We prove a conjecture of K. Schmidt in algebraic dynamical system theory onthe growth of the number of components of fixed point sets. We also prove arelated conjecture of Silver and Williams on the growth of homology torsions offinite abelian covering of link complements. In both cases, the growth isexpressed by the Mahler measure of the first non-zero Alexander polynomial ofthe corresponding modules. In the case of non-ablian covering, the growth of torsion is less thanor equal to the hyperbolic volume (or Gromov norm) of the knot complement.

(joint work with M. Macasieb) Let $K$ be a hyperbolic $(-2, 3, n)$ pretzel knot and $M = S^3 \setminus K$ its complement. For these knots, we verify a conjecture of Reid and Walsh: there are at most three knotcomplements in the commensurability class of $M$. Indeed, if $n \neq 7$, weshow that $M$ is the unique knot complement in its class.