On the conditioning of the Padé map and related questions

Padé approximants play an important role in signal processing, sparse interpolation and exponential analysis. In this talk we will report about recent results concerning the forward and backward conditioning of the (real) Padé map, which sends a vector of Taylor coefficients onto the normalized vector of coefficients of the Padé numerator and denominator. In particular, we show that this map is not necessarily well conditioned for robust Padé approximants in the sense of Trefethen et al.

We will also discuss the condition number of related non-linear maps.