Abstract: Over three centuries, the approximation of functions by linear superposition has evolved from polynomials to wavelet Fourier series. We will survey some of these methods and examine how desirable features beyond ease of computation have been introduced. We will see how Haar and Daubechies developed special bases with prescribed features that could be adapted to particular approximation problems. Some applications include image compression and de-noising.
This will be a hybrid seminar. All are very welcome to join either online or in person. Venue: Luigi Stasi Seminar Room (ICTP Leonardo Da Vinci Building), for those wishing to attend in person.