AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

A Unified Framework for Diffusion Model Unlearning with f-Divergence

Nicola Novello, Federico Fontana, Luigi Cinque, Deniz Gunduz, Andrea M Tonello

ICML 2026 · Image generation › Text-to-image diffusion · approach: Concept erasure in diffusion models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace MSE (KL) loss with closed-form Hellinger loss (ablation)diffusion unlearning (CAbl-style) · object and style erasure, SD 1.4/2.1prior preservation and KID: KID 0.250 -> 0.005 (Grumpy Cat, erasing Dali)Improved
replace MSE with Hellinger loss (ablation)diffusion unlearning · nudity erasure I2P, SD 1.5nudity generation rate: 0.120 -> 0.063Improved
replace MSE with Hellinger loss (ablation)diffusion unlearning · Ring-A-Bell adversarial prompts, SD 1.5nudity generation rate: 0.063 -> 0.157Hurt
replace MSE with chi2 loss (ablation)diffusion unlearning · object and style erasureerasure strength vs preservationMixed
replace MSE with chi2 loss (ablation)diffusion unlearning · nudity erasure I2P, SD 1.5nudity generation rate: 0.120 -> 0.204Hurt
use Hellinger instead of MSE for sequential erasure of 10 styles (ablation)diffusion unlearning · 10 artistic styles, SD 1.4prior preservation and erasureImproved
use Hellinger loss with bounded gradients (ablation)diffusion unlearning · fine-tuning stabilitygeneration stabilityImproved
use variational min-max loss instead of closed-form (ablation)Hellinger f-DMU · Van Gogh erasure, SD 2.1KID (preserved): 0.027 -> 0.051Hurt
use variational loss instead of closed-form (ablation)f-DMU · unlearning speediterations to low CS and CAImproved
use Hellinger variational loss instead of DoCovariational f-DMU · Van Gogh erasure, SD 2.1preserved CLIP accuracy and erasureImproved
use Hellinger variational loss instead of Hellinger closed-form for nudity (ablation)f-DMU · I2P, SD 1.5nudity generation rate: 0.063 -> 0.430Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy