7th Symposium On Conformal & Probabilistic Prediction With Applications (COPA 2018)

About the Symposium

The purpose of Conformal Prediction is to complement predictions delivered by various algorithms of Machine Learning with valid probability measures of their accuracy and reliability. Conformal Prediction is a universal tool in the sense that it can be used in combination with any known algorithm, such as SVM, Neural Networks, etc.

A sister method of Venn prediction was developed at the same time as conformal prediction and is used for probabilistic prediction. The COPA series of workshops/symposia is a home for work in both conformal and Venn prediction. The aim of this symposium is to serve as a forum for the presentation of new and ongoing work and the exchange of ideas between researchers on any aspect of Conformal and Probabilistic Prediction and their applications to interesting problems of any field.

University of London Kolmogorov Lecture

Prof. Vladimir Vapnik will deliver the 2018 University of London Kolmogorov Lecture: "Rethinking Statistical Learning Theory: Learning Using Statistical Invariants."

Abstract

In the talk I will considers Teacher-Student interaction in learning processes. I will introduce a new learning paradigm, called Learning Using Statistical Invariants (LUSI), which is different from the classical one.

In the classical paradigm, learning machine constructs, using data, a classification or regression function that minimizes the expected loss; it is thus data-driven learning. In the LUSI paradigm, in order to construct the desired classification or regression function using both data and Teacher's input, learning machine computes statistical invariants that are specific for the problem, and then minimizes the expected loss in a way that preserves these invariants; so it is both data- and intelligence-driven learning.

From a mathematical point of view, methods of the classical paradigm employ mechanisms of strong convergence of approximations to the desired function, whereas methods of the new paradigm employ both strong and weak convergence mechanisms. This can significantly increase the rate of convergence.