[KongchangAI]
ConceptDiffusion Guidance

Diffusion Guidance

A technique that injects additional gradients or biases into the denoising network during the sampling stage of a diffusion model, steering generated samples toward a target distribution without retraining the model.

Timeline (last 90 days)

Nov 10

At sampling time, diffusion guidance shifts each emulated climate model distribution toward observations, with its strength selected through perfect-model experiments.

Unverified50%
Nov 10

On seventeen synthetic targets, guidance more than halves the median bias of the 1st and 99th percentiles of daily temperature, improving probability scores and reliability.

Unverified50%
Nov 10

Under three emissions scenarios, guidance reduces the median RMSE of global mean temperature by 32-57%.

Unverified50%
Nov 10

In detailed benchmarks, guidance outperforms existing bias correction methods on daily field probability scores.

Unverified50%
Nov 10

Diffusion guidance injects additional gradients or biases into the denoising network during the sampling stage to steer generated samples toward a target distribution without retraining the model.

Unverified50%

All Facts (5)

Source Articles