search

tstandley / taskgrouping

Code for Which Tasks Should Be Learned Together in Multi-task Learning?
Other
93 stars 14 forks source link

Question about "optimal network choice" #12

Open ShuGuoJ opened 1 year ago

ShuGuoJ commented 1 year ago

In legends of figure 2, 3, 4, 5, and 6, ESA refers to the technology of early stopping approximation, and HOA refers to the technology of higher order approximation, so what does the "optimal network choice" refer to?

tstandley commented 1 year ago

Thanks for your interest!

Optimal network choice is what you get when you train all networks to convergence so you know exactly how they will perform and then use an algorithm like the one in the appendix to choose the best combination of networks.

No approximation strategy can find a better grouping.

On Tue, Oct 10, 2023, 6:07 AM ShuGuoJ @.***> wrote:

In legends of figure 2, 3, 4, 5, and 6, ESA refers to the technology of early stopping approximation, and HOA refers to the technology of higher order approximation, so what does the "optimal network choice" refer to?

— Reply to this email directly, view it on GitHub https://github.com/tstandley/taskgrouping/issues/12, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEDABXVOAGMK4FT4EMXGLGTX6VCC5AVCNFSM6AAAAAA52KWGCSVHI2DSMVQWIX3LMV43ASLTON2WKOZRHEZTKMZTGY4TQOA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

ShuGuoJ commented 1 year ago

Thanks for your timely reply!