Metraipolis › archive › dc_a41932ba8a8e
Gaussian Core LoRA: Distribution-Aware Dynamic Adaptation for Broad Concept Erasure — arXiv:2609.01433
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.01433
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T07:13:57.168Z
- SHA-256 of the archived text
96e15ae3fc313ec7f35944854b8786a79314ad2e21e8e8df9af7ca9c4446292b- Stable URL of this page
https://metraipolis.com/documento/dc_a41932ba8a8e
How to cite
Gaussian Core LoRA: Distribution-Aware Dynamic Adaptation for Broad Concept Erasure — arXiv:2609.01433 — arXiv (en), https://arxiv.org/abs/2609.01433 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a41932ba8a8e (Metraipolis archive dc_a41932ba8a8e, sha256 96e15ae3fc313ec7f35944854b8786a79314ad2e21e8e8df9af7ca9c4446292b)
BibTeX
@misc{dca41932ba8a8e,
title = {Gaussian Core LoRA: Distribution-Aware Dynamic Adaptation for Broad Concept Erasure — arXiv:2609.01433},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.01433},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a41932ba8a8e; sha256 96e15ae3fc313ec7f35944854b8786a79314ad2e21e8e8df9af7ca9c4446292b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01433. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Concept erasure aims to suppress unsafe, privacy-sensitive, or undesirable generations in text-to-image diffusion models while preserving benign semantics, visual quality, and deployment efficiency. Existing adapter-based methods, such as Low-Rank Adaptation (LoRA), typically freeze the diffusion backbone and learn lightweight parameter updates to steer generation away from target semantics. However, these methods usually assign a static semantic erasure direction to each target concept. This assumption is overly coarse for broad and complex target concepts, since a concept often contains multiple latent semantic prototypes involving different objects, scenes, or relations, and requires different local erasure directions. A single LoRA update averages these heterogeneous erasure demands, leading to under-erasure on difficult prototypes and over-editing of nearby benign semantics. To address this limitation, we propose Gaussian Core LoRA, a distribution-aware low-rank adaptation framework. It fits a Gaussian mixture model in the prompt feature space to estimate latent semantic prototypes within the target concept. During inference, each input prompt is projected into this feature space to compute its Gaussian posterior responsibilities, which condition the core generator to produce a prompt-specific, norm-bounded residual reconfiguration of the shared LoRA rank space. This enables prototype-adaptive erasure with a single lightweight adapter. Compared with the strongest baseline on each metric, Gaussian Core LoRA reduces average Attack Success Rate (ASR) by 7.95%, lowers COCO Fr'echet Inception Distance (FID) by 14.72%, and improves CLIP Score by 4.98%. Further experiments show robustness to adversarial prompts, scalability to multi-identity and multi-style erasure, and compatibility with SDXL and FLUX. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-02 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.01433 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T07:00:58.990Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).