In one calculation, adjoint sensitivity analysis provides the gradient of a
quantity of interest with respect to all system's parameters. Conventionally,
adjoint solvers need to be implemented by differentiating computational models,
which can be a cumbersome task and is code-specific.
Symplectic Adjoint Guidance (SAG) proposes a training-free guided sampling approach in diffusion models, improving accuracy and quality of image and video generation by estimating the clean image using multiple function calls and obtaining gradients efficiently.