English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Fine-tuning diffusion models through personalized datasets is an acknowledgedmethod for improving generation quality across downstream tasks, which,however, often inadvertently generates unintended concepts such as watermarksand QR codes, attributed to the limitations in image sources and collectingmethods within specific downstream tasks. Existing solutions suffer fromeliminating these unintentionally learned implicit concepts, primarily due tothe dependency on the model's ability to recognize concepts that it actuallycannot discern. In this work, we introduce Geom-Erasing, a novel approach thatsuccessfully removes the implicit concepts with either an additional accessibleclassifier or detector model to encode geometric information of these conceptsinto text domain. Moreover, we propose Implicit Concept, a novel image-textdataset imbued with three implicit concepts (i.e., watermarks, QR codes, andtext) for training and evaluation. Experimental results demonstrate thatGeom-Erasing not only identifies but also proficiently eradicates implicitconcepts, revealing a significant improvement over the existing methods. Theintegration of geometric information marks a substantial progression in theprecise removal of implicit concepts in diffusion models.

Geom-Erasing: Geometry-Driven Removal of Implicit Concept in Diffusion Models