Reaction
A possible way to speed up synchronous RL using MOPD for branch-train-merge
A user calls the algorithm idea “incredibly cursed” but does not explain how it would work.
TLDR
A user says they thought of an “incredibly cursed” algorithm that could speed up synchronous reinforcement learning using MOPD for branch-train-merge. The post gives no implementation details or results.
Combined views
—
1 Source, first seen ago
— likes— comments— saves— reposts