Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.18767 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917353180102656 |
|---|---|
| author | Pham, Duc Hao Truong, Van Duy Dinh, Duy Khanh Nguyen, Tien Cuong Ngo, Dien Hy Bui, Tuan Anh |
| author_facet | Pham, Duc Hao Truong, Van Duy Dinh, Duy Khanh Nguyen, Tien Cuong Ngo, Dien Hy Bui, Tuan Anh |
| contents | Concept unlearning has emerged as a promising direction for reducing the risks of harmful content generation in text-to-image diffusion models by selectively erasing undesirable concepts from a model's parameters. Existing approaches typically rely on keywords to identify the target concept to be unlearned. However, we show that this keyword-based formulation is inherently limited: a visual concept is multi-dimensional, can be expressed in diverse textual forms, and often overlap with related concepts in the latent space, making keyword-only unlearning, which imprecisely indicate the target concept is brittle and prone to over-forgetting. This occurs because a single keyword represents only a narrow point estimate of the concept, failing to cover its full semantic distribution and entangled variations in the latent space. To address this limitation, we propose Diversified Unlearning, a distributional framework that represents a concept through a set of contextually diverse prompts rather than a single keyword. This richer representation enables more precise and robust unlearning. Through extensive experiments across multiple benchmarks and state-of-the-art baselines, we demonstrate that integrating Diversified Unlearning as an add-on component into existing unlearning pipelines consistently achieves stronger erasure, better retention of unrelated concepts, and improved robustness against adversarial recovery attacks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_18767 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Concept is More Than a Word: Diversified Unlearning in Text-to-Image Diffusion Models Pham, Duc Hao Truong, Van Duy Dinh, Duy Khanh Nguyen, Tien Cuong Ngo, Dien Hy Bui, Tuan Anh Artificial Intelligence Concept unlearning has emerged as a promising direction for reducing the risks of harmful content generation in text-to-image diffusion models by selectively erasing undesirable concepts from a model's parameters. Existing approaches typically rely on keywords to identify the target concept to be unlearned. However, we show that this keyword-based formulation is inherently limited: a visual concept is multi-dimensional, can be expressed in diverse textual forms, and often overlap with related concepts in the latent space, making keyword-only unlearning, which imprecisely indicate the target concept is brittle and prone to over-forgetting. This occurs because a single keyword represents only a narrow point estimate of the concept, failing to cover its full semantic distribution and entangled variations in the latent space. To address this limitation, we propose Diversified Unlearning, a distributional framework that represents a concept through a set of contextually diverse prompts rather than a single keyword. This richer representation enables more precise and robust unlearning. Through extensive experiments across multiple benchmarks and state-of-the-art baselines, we demonstrate that integrating Diversified Unlearning as an add-on component into existing unlearning pipelines consistently achieves stronger erasure, better retention of unrelated concepts, and improved robustness against adversarial recovery attacks. |
| title | A Concept is More Than a Word: Diversified Unlearning in Text-to-Image Diffusion Models |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2603.18767 |