Over the past two years, a stream of breakthrough works analyzing the continuum limits of sampling and optimization algorithms have solidified the importance of PDEs for both understanding existing algorithms and developing new ones. On one hand, the PDE perspective brings with it a mature theory of numerical analysis, leading to new discretizations of existing algorithms. Due to historical connections with physics, PDE also can provide a range of new energy landscapes on which to build algorithms. On the other hand, in the PDE community, there is an acute need for improved algorithms in numerical simulations — indeed, efficient sampling is essential to particle methods and recent advances in optimization have enabled the discovery of new numerical methods built upon a PDE’s variational structure. The goal of this workshop is to nurture the development of a continuum mathematical framework that can unite various classes of discrete algorithms in PDEs, sampling, and optimization.
All events take place in the Calvin Lab auditorium.
Further details about this workshop will be posted in due course. Enquiries may be sent to the organizersworkshop-gm2 [at] lists.simons.berkeley.edu ( at this address.)