18821611. TRAJECTORY STITCHING FOR ACCELERATING DIFFUSION MODELS (NVIDIA Corporation)
TRAJECTORY STITCHING FOR ACCELERATING DIFFUSION MODELS
Organization Name
Inventor(s)
De-An Huang of Cupertino CA US
Anima Anandkumar of Pasadena CA US
TRAJECTORY STITCHING FOR ACCELERATING DIFFUSION MODELS
This abstract first appeared for US patent application 18821611 titled 'TRAJECTORY STITCHING FOR ACCELERATING DIFFUSION MODELS
Original Abstract Submitted
Diffusion models are machine learning algorithms that are uniquely trained to generate high-quality data from an input lower-quality data. Diffusion probabilistic models use discrete-time random processes or continuous-time stochastic differential equations (SDEs) that learn to gradually remove the noise added to the data points. With diffusion probabilistic models, high quality output currently requires sampling from a large diffusion probabilistic model which corners at a high computational cost. The present disclosure stitches together the trajectory of two or more inferior diffusion probabilistic models during a denoising process, which can in turn accelerate the denoising process by avoiding use of only a single large diffusion probabilistic model.