Speaker:  Michele Pagani


Quando:  05/06/2020  15:00


Dove:  videoconferenza online


Abstract
Venerdì 5 giugno 2020 alle ore 15:00, il prof. Michele Pagani dell'Université Paris 7, presenterà il seminario di Logica e Informatica Teorica dal titolo: "Automatic differentiation in PCF". Abstract: Backpropagation is a classic automatic differentiation algorithm computing the gradient of functions specified by a certain class of simple, firstorder programs, called computational graphs. It is a fundamental tool in several fields, most notably machine learning, where it is the key for efficiently training (deep) neural networks. Recent years have witnessed the quick growth of a research field called differentiable programming, the aim of which is to express computational graphs more synthetically and modularly by resorting to actual programming languages endowed with control flow operators and higherorder combinators, such as map and fold. We extend the backpropagation algorithm to a paradigmatic example of such a programming language: we define a compositional program transformation from PCF (a Turing complete simplytyped lambdacalculus) to itself augmented with a notion of linear negation, and prove that this computes almost everywhere the gradient of the source program with the same efficiency as firstorder backpropagation. The transformation is completely effectfree and thus provides a purely logical understanding of the dynamics of backpropagation. Per partecipare al seminario, richiedere il link all’indirizzo email vitomichele.abrusci@uniroma3.it o cliccare sul seguente link Teams Meeting 