Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
This 17-minute conference talk from the LAFI 2025 workshop presents research on data-parallel differentiation through optic composition by Paul Wilson and Fabio Zanasi from the University of Southampton and University College London. Explore an innovative algorithm that transforms morphisms in symmetric monoidal categories into their reverse derivatives using a data-parallel approach. Learn how this algorithm achieves logarithmic time complexity on PRAM machines and linear time on sequential machines. The presentation demonstrates how to compute reverse derivatives efficiently when given a symmetric monoidal category with generators, operations, and derivative choices for each operation. This research was presented at the LAFI workshop on January 19, 2025, as part of the POPL25 conference sponsored by ACM SIGPLAN.
