We are adding exact first-class derivative calculation operators
(Automatic Differentiation or AD) to the lambda calculus, and
embodying the combination in a production-quality fast system suitable
for numeric computing in general, and compositional machine learning
methods in particular. To the programming language community, we seek
to contribute a way to make numeric software faster, more robust, and
easier to write.
To the machine learning community, in addition to the above practical
benefits, we seek to contribute a system that embodies
*compositionality*, in that gradient optimisation can be automatically
and efficiently performed on systems themselves consisting of many
components, even when such components may internally perform
optimisation. (Examples of this include, say, optimisation of the
rules of a multi-player game to cause the players actions to satisfy
some desiderata, where the players themselves optimise their own
strategy with using a simple model of the opponent which they optimise
according to their opponent's behaviour; or multi-agent learning where
one agent learns an internal model of another agent, where that
internal model itself performs learning.)
To this end, we are seeking two postdoctoral researchers and one
research programmer with interest and experience in a cohert subset
of: programming language theory, numerics, automatic differentiation,
and machine learning.
Inquiries to: Barak A. Pearlmutter