Dense Reward
A reward form in reinforcement learning that provides intermediate feedback at every state or step. In contrast to sparse rewards, it can substantially shorten the exploration path.
Timeline (last 90 days)
Dense rewards provide continuous feedback at each step or state, which can efficiently guide policy learning, but may also over-constrain the policy near the reference trajectory.
The scene graph provides dense rewards for each task stage, ensuring that when the policy starts from a reset state, the exploration path is sufficiently short and clearly guided.
All Facts (2)
Dense rewards provide continuous feedback at each step or state, which can efficiently guide policy learning, but may also over-constrain the policy near the reference trajectory.
50%UnverifiedThe scene graph provides dense rewards for each task stage, ensuring that when the policy starts from a reset state, the exploration path is sufficiently short and clearly guided.
50%