[Pl-seminar] 1/28 Seminar: Damiano Mazza: "Automatic Differentiation in PCF"

[Nathaniel Yazdani]

Reminder that this is tomorrow!

> On Jan 23, 2020, at 8:10 PM, Nathaniel Yazdani <yazdani.n at husky.neu.edu> wrote:
> 
> NUPRL Seminar Presents
> 
> Damiano Mazza
> CNRS, Université Sorbonne Paris Nord
> 
> 10:00AM to 11:30AM
> Tuesday, January 28th, 2020
> Room 366 WVH (www.ccs.neu.edu/home/wand/directions.html)
> 
> Automatic Differentiation in PCF
> 
> Abstract
> 
> Automatic differentiation (AD) is the science of efficiently computing
> the derivative (or gradient, or Jacobian) of functions specified by
> computer programs. It is a fundamental tool in several fields, most
> notably machine learning, where it is the key for training deep neural
> networks. Albeit AD techniques natively focus on a restricted class of
> programs, namely first-order straight-line programs, the rise of
> so-called differentiable programming in recent years has urged for the
> need of applying AD to complex programs, endowed with control flow
> operators and higher-order combinators, such as map and fold. In this
> talk, I will discuss the extension of AD algorithms to PCF, a(n idealized)
> purely functional programming language. We will first consider the
> simply-typed lambda-calculus, showing in particular how linear
> negation is related to reverse-mode AD (a.k.a., backpropagation), and
> then see how the extra features of PCF, namely full recursion and
> conditionals, may be dealt with, stressing the difficulties posed by the
> latter.
> 
> Joint work with Aloïs Brunel (Deepomatic) and Michele Pagani (IRIF,
> Université de Paris).
> 
> Bio
> 
> Damiano Mazza is chargé de recherche (researcher) at CNRS, working at
> the CS lab of Université Sorbonne Paris Nord, a position he has held
> since 2008. He obtained his Ph.D. in 2006 at the Institut de
> Mathématiques de Luminy, in Marseille. Before that, he studied CS
> Engineering in Rome, Italy, which is where he is from. His research
> interest are at the interface between logic in computer science, the
> theory of programming languages and complexity theory, with a bias
> towards the viewpoint provided by linear logic and category theory.
>