You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi authors,
I fail to understand why the randomly sampled epsilons are dependent in respect to the timestep axis. And likewise, I fail to understand why does epsilon_hat is independent. Thank you for the great work and eager for your reply!
The text was updated successfully, but these errors were encountered:
Eq1 establishes a connection between x_t and x_{t-1}, incorporating independent noise (eta). Upon rephrasing it as Eq2, we depict a connection from x_0 to x_t leading to a noise (epsilon) that takes into account all preceding etas from Eq 1.
Consequently, the epsilon linking x_0 to x_{t-1} demonstrates a significant correlation with the epsilon linking x_0 to x_t. These epsilons lack independence.
However, in our scenario, the sampling process generating x_t was devised as a separate auxiliary process between all timesteps. There exists no connection between the creation of x_t and the creation of x_{t-1}. This suggests that the epsilons creating x_t are independent across all timesteps.
Hi authors,
I fail to understand why the randomly sampled epsilons are dependent in respect to the timestep axis. And likewise, I fail to understand why does epsilon_hat is independent. Thank you for the great work and eager for your reply!
The text was updated successfully, but these errors were encountered: