Jacob Kelly
@jacobjinkelly
This is pretty RAD
Ryan Adams@ryan_p_adams · Jul 24, 2020Automatic differentiation collapses the linearized computational graph to compute a Jacobian. We don’t need exact gradients for SGD, so let’s use cheap Monte Carlo estimators instead: arxiv.org/abs/2007.10412
github.com/PrincetonLIPS/…
@denizzokt @AlexBeatson @NMcGreivy @jaduol1
github.com/PrincetonLIPS/…
@denizzokt @AlexBeatson @NMcGreivy @jaduol1
Open quoted post →
0 6