our next and last One World ABC seminar of the season is quickly approaching. 

On Thursday June 23, at 1.30pm UK time, Cosma Shalizi will talk about "Matching Random Features", with an abstract reported below.

https://bristol-ac-uk.zoom.us/j/99993089580?pwd=NUtzT0VLdGNvb0tJMFRNMXl6SlNsQT09  

Password: 486380

We are looking forward to seeing you on Thursday,

best wishes,

Massimiliano on the behalf of the One World ABC Seminar Organisers 

Speaker: Cosma Shalizi, Carnegie Mellon University

Title: Matching Random Features

Abstract: We can do statistical inference on simulation models by adjusting the parameters in the simulation so that the values of randomly chosen functions of the simulation output match the values of those same functions calculated on the data. Results from the ”statespace reconstruction” or “geometry from a time series” literature in nonlinear dynamics indicate that just 2d + 1 such functions will typically suffice to identify a model with a d-dimensional parameter space. Results from the “random features” literature in machine learning suggest that using random functions of the data can be an efficient replacement for using optimal functions. In this talk, I sketch some of the key results, present numerical evidence about the new method’s properties, and lay out an agenda for research.