Research map / Image generation / Text-to-image diffusion
Improving Geo-diversity of Generated Images with Contextualized Vendi Score Guidance
Reyhane Askari Hemmat, Melissa Hall, Alicia Yi Sun, Candace Ross, Michal Drozdzal, Adriana Romero-Soriano
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add Vendi Score Guidance (memory bank) to the LDM | LDM · GeoDE | F1 (avg): 0.364 -> 0.399 | Improved |
| add exemplar-image contextualization to VSG (ablation) | VSG · GeoDE | F1 (avg): 0.399 -> 0.455 | Improved |
| add exemplar-image contextualization to VSG (ablation) | VSG · DollarStreet | F1 (avg): 0.457 -> 0.497 | Improved |
| add contextualization to VSG (ablation) | VSG · GeoDE | CLIPScore (avg): 0.180 -> 0.254 | Improved |
| add contextualization to VSG (ablation) | VSG · DollarStreet | Recall (avg): 0.516 -> 0.511 | Hurt |
| add memory-bank Vendi guidance (alpha) on top of exemplar-only guidance (ablation) | c-VSG · GeoDE | Recall (avg): 0.464 -> 0.493 | Improved |
| replace Vendi-based guidance with Synonyms prompting | LDM · GeoDE | F1 (avg): 0.364 -> 0.357 | Hurt |
| increase exemplar weight beta from alpha to 4 alpha (ablation) | c-VSG · GeoDE | Recall (avg): 0.547 -> 0.472 | Hurt |
| increase exemplar weight beta from alpha to 4 alpha (ablation) | c-VSG · GeoDE | Precision (avg): 0.384 -> 0.428 | Improved |
| increase number of exemplar images per object from 2 to 20 (ablation) | c-VSG · GeoDE | F1 (worst region): 0.444 -> 0.433 | No clear effect |
| use region-stratified exemplars instead of random exemplars (ablation) | c-VSG · GeoDE | Recall (worst region): 0.438 -> 0.459 | Improved |
| add c-VSG to LDM versus Semantic Guidance baseline | LDM · GeoDE | Precision (avg): 0.459 vs 0.424 | Mixed |
Closest papers by meaning
- Image Generation Diversity Issues and How to Tame Them (CVPR 2025 · similarity 0.89)
- No Filter: Cultural and Socioeconomic Diversity in Contrastive Vision-Language Models (NeurIPS 2024 · similarity 0.88)
- Increasing the Utility of Synthetic Images through Chamfer Guidance (NeurIPS 2025 · similarity 0.88)
- Multilingual Diversity Improves Vision-Language Representations (NeurIPS 2024 · similarity 0.87)
- RA-ClipScore: Making Generative Model Evaluation More Interpretable (NeurIPS 2026 · similarity 0.87)
- MMM-RS: A Multi-modal, Multi-GSD, Multi-scene Remote Sensing Dataset and Benchmark for Text-to-Image Generation (NeurIPS 2024 · similarity 0.86)
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.
