Beyond Overcorrection: Evaluating Diversity in T2I Models with DivBench
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913935629746176 |
|---|---|
| author | Friedrich, Felix Welsch, Thiemo Ganesha Brack, Manuel Schramowski, Patrick Kersting, Kristian |
| author_facet | Friedrich, Felix Welsch, Thiemo Ganesha Brack, Manuel Schramowski, Patrick Kersting, Kristian |
| contents | Current diversification strategies for text-to-image (T2I) models often ignore contextual appropriateness, leading to over-diversification where demographic attributes are modified even when explicitly specified in prompts. This paper introduces DIVBENCH, a benchmark and evaluation framework for measuring both under- and over-diversification in T2I generation. Through systematic evaluation of state-of-the-art T2I models, we find that while most models exhibit limited diversity, many diversification approaches overcorrect by inappropriately altering contextually-specified attributes. We demonstrate that context-aware methods, particularly LLM-guided FairDiffusion and prompt rewriting, can already effectively address under-diversity while avoiding over-diversification, achieving a better balance between representation and semantic fidelity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_03015 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Beyond Overcorrection: Evaluating Diversity in T2I Models with DivBench Friedrich, Felix Welsch, Thiemo Ganesha Brack, Manuel Schramowski, Patrick Kersting, Kristian Computation and Language Computers and Society Machine Learning Current diversification strategies for text-to-image (T2I) models often ignore contextual appropriateness, leading to over-diversification where demographic attributes are modified even when explicitly specified in prompts. This paper introduces DIVBENCH, a benchmark and evaluation framework for measuring both under- and over-diversification in T2I generation. Through systematic evaluation of state-of-the-art T2I models, we find that while most models exhibit limited diversity, many diversification approaches overcorrect by inappropriately altering contextually-specified attributes. We demonstrate that context-aware methods, particularly LLM-guided FairDiffusion and prompt rewriting, can already effectively address under-diversity while avoiding over-diversification, achieving a better balance between representation and semantic fidelity. |
| title | Beyond Overcorrection: Evaluating Diversity in T2I Models with DivBench |
| topic | Computation and Language Computers and Society Machine Learning |
| url | https://arxiv.org/abs/2507.03015 |