Table 2:

Optimal Time-Dependent Learning Rates.

SyntheticMediamill
Bio-CCA $(η0,γ,τ)$ $(10-3,10-4,0.1)$ $(10-2,10-4,0.1)$
Gen-Oja $(β0,γ)$ $(1,10-2)$ $(10-2,10-3)$
Asym-NN $(η0,α)$ $(2×10-4,5×10-6)$ $(2×10-3,5×10-6)$
Bio-RRR $(η0,γ,μ)$ $(10-3,10-4,10)$ $(10-2,10-5,10)$
Adaptive Bio-CCA $(η0,γ,τ)$ $(10-3,10-4,0.1)$ $(10-2,10-4,0.5)$
SyntheticMediamill
Bio-CCA $(η0,γ,τ)$ $(10-3,10-4,0.1)$ $(10-2,10-4,0.1)$
Gen-Oja $(β0,γ)$ $(1,10-2)$ $(10-2,10-3)$
Asym-NN $(η0,α)$ $(2×10-4,5×10-6)$ $(2×10-3,5×10-6)$
Bio-RRR $(η0,γ,μ)$ $(10-3,10-4,10)$ $(10-2,10-5,10)$
Adaptive Bio-CCA $(η0,γ,τ)$ $(10-3,10-4,0.1)$ $(10-2,10-4,0.5)$
Close Modal