Equilibrium Discovery in Modular Deep Learning Architectures
Carlos E. Perez
115

I’m going to have to disagree here. Markets historically and consistently have underassigned economic value to high net value contributors such as mathematicians. They don’t necessarily lead to socially optimal outcomes and nash equilibria in market systems makes things even worse.

I don’t think basing deep learning off these principles will help with anything and is far from “principled”. As nassim-taleb said, “macroeconomic analysis works worse than astrology”. In other words, using market systems as an ensemble strategy is not principled at all.

TL;DR: Follow the mathematicians for ensembling, not the economists.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.