Implicit Function




and Dynamics

February 19–22, 2020

Schloß Rauischholzhausen

The winter school will present two incarnations of the implicit function theorem to master and PhD students: the implicit function theorem of Nash and Moser, and the implicit function theorem in scale calculus and polyfold-theory as developed by Hofer, Wysocki and Zehnder.

There will be two mini-courses on the theoretical framework of each incarnation which will be complemented by talks on classical and current applications to analysis and geometry.

Lecture courses

Massimiliano Berti

Nash-Moser implicit function theorems and applications in dynamics

In these lectures I will prove an abstract Nash-Moser implicit function theorem for a nonlinear operator acting in scales of Banach spaces. In contrast to the usual implicit function theorem the key feature is that the linearized operator admits and approximate inverse which in unbounded in a fixed Banach space. This abstract theorem is taylored to prove, in particular, the classical KAM (Kolmogorov-Arnold-Moser) result concerning the persistence of quasi-periodic solutions for sufficiently differentiable nearly integrable Hamiltonian systems. The proof of this result will be given in the talk by F. Giuliani. All the functional analytic preliminaries of smoothing operators, tame estimates and interpolation inequalities will also be proved.

Zhengyi Zhou

Implicit function theorems in scale-calculus and polyfold theory

Implicit function theorems in scale-calculus and more generally polyfold theory are at the heart of the applications to symplectic topology for which these theories were developed. At the same time, the notion of scale-derivative does not require continuity of the scale-differential in the operator topology as the base point varies, which prevents the obvious analogues in scale-calculus of classical implicit function theorems for Fréchet differentiable maps from holding. Much effort has gone into understanding under what conditions an implicit function theorem does hold, miraculously with success in all cases that are known to be relevant to symplectic topology. Hofer-Wysocki-Zehnder formulated the notion of a scale-Fredholm map and proved the implicit function theorem for scale-Fredholm sections of polyfold bundles. In joint work with Benjamin Filippenko and Katrin Wehrheim, we explore the boundary between success and failure of the implicit function theorem in scale-calculus, along the way justifying the assumption of a contraction germ property in the definition of scale-Fredholm. In this four lecture course, I will tell this story starting from the definitions of scale-calculus.

Additional talks

Alberto Abbondandolo



Pietro Baldi



Lucas Dahinden



Filippo Giuliani

Proof of the KAM theorem by a Nash-Moser approach

In this talk we show a proof of the classical KAM theorem in the finite differentiable case by applying the Nash-Moser implicit function theorem presented by M. Berti in his course. More precisely, we prove the existence of Lagrangian invariant tori supporting quasi-periodic motions for small C^k-smooth perturbations (with k large) of a family of isochronous Hamiltonian systems.

Colin Guillarmou

Marked length spectrum local rigidity

We will explain how to show a local rigidity result for the marked length spectrum of Riemannian metrics with negative curvature, using a weak version of implicit function theorem.

Irene Seifert

Finding periodic delay orbits via the Polyfold-IFT

A differential delay equation in R^n is of the form \dot{x}(t) = X_{t}(x(t-\tau)) with a vector field X on R^n and a delay parameter \tau. If there is a (suitably non-degenerate) periodic solution x without delay, intuitively one would expect existence of periodic solutions for sufficiently small positive delay. However, classical implicit function theorems fail to imply this, since the equation above is not smooth in the delay parameter. In my talk, I will show how to use the polyfold-IFT to overcome this problem in a natural set-up.​

Thibault Lefeuvre

Local rigidity of the boundary distance function

On a Riemannian manifold with boundary, the boundary distance function is the map assigning to each pair of points on the boundary, the Riemannian distance between the two points. It is conjectured that under some mild assumptions on the metric and/or on the topology of the manifold (that I will detail), one can reconstruct the metric up to isometries. This theoretical question naturally arises in seismic tomography, where one tries to reconstruct the inner structure of the crust from the knowledge of travel times of seismic P- and S-waves from the earthquake to seismic stations. I will explain some partial results on this question, involving (in two different steps) an implicit function theorem. Some tools will be reinvested in a subsequent talk by C. Guillarmou in order to deal with a rather similar problem on closed manifolds.

Dingyu Yang



Practical information


Schloßpark 1

35085 Ebsdorfergrund



February, 19 – before lunch


February, 22 – after lunch


Please send an e-mail to Petra Kuhl.


February 19, before lunch

We will organise shuttle buses that will bring you up to the castle. The meeting point for the shuttle is in front of the main station Marburg (Lahn).


February 22, after lunch

We will drive together at 1:45 pm with a bus to the main train station Marburg (Lahn), which will be reached by 2:20 pm at the latest.


Lodging and meals will be provided by the castle. All local costs are covered by us. All lectures will take place at the castle.