Research map / Image generation / Text-to-image diffusion
RankT2I: A Submodular Framework for Discovering Interpretable and Diverse Semantics in Text-to-Image Models
Ritika Allada, Pinar Yanardag
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 |
|---|---|---|---|
| rank semantics with submodular objective instead of SliderSpace directions (FLUX) | RankT2I · semantic discovery | TCE: 12.5769 -> 22.8102 | Improved |
| rank semantics with submodular objective instead of NoiseCLR (SD) | RankT2I · semantic discovery | VQAScore: 0.5484 -> 0.8177 | Improved |
| training-free ranking instead of trained directions (SD, SliderSpace) | RankT2I · time to find 100 semantics | total time: 391m 48s -> 43m 14s | Improved |
| RankT2I vs K-Means attribute retrieval (SD) | RankT2I · attribute retrieval | I2I CLIP similarity: 0.8433 vs 0.8436 | No clear effect |
| RankT2I vs cosine-similarity retrieval (FLUX) | RankT2I · attribute retrieval | CLIP-T: 0.2920 -> 0.2799 | Mixed |
| increase diversity weight lambda (ablation) | RankT2I · top-25 attributes | number of categories | Improved |
| increase relevance weight alpha or editability weight beta (ablation) | RankT2I · top-25 attributes | number of categories | Hurt |
| rank by diversity only (ablation) | RankT2I · dress domain FLUX | editability | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SliderSpace — SliderSpace: Decomposing the Visual Capabilities of Diffusion Models (ICCV 2025) · better on 8 of 8 shared results
- NoiseCLR — NoiseCLR: A Contrastive Learning Approach for Unsupervised Discovery of Interpretable Directions in Diffusion Models (CVPR 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Localizing and Editing Knowledge In Text-to-Image Generative Models (ICLR 2024 · similarity 0.89)
- PairEdit: Learning Semantic Variations for Exemplar-based Image Editing (NeurIPS 2025 · similarity 0.88)
- The Narrow Gate: Localized Image-Text Communication in Native Multimodal Models (NeurIPS 2025 · similarity 0.88)
- MARS: Mixture of Auto-Regressive Models for Fine-grained Text-to-image Synthesis (AAAI 2025 · similarity 0.88)
- VLKEB: A Large Vision-Language Model Knowledge Editing Benchmark (NeurIPS 2024 · similarity 0.88)
- DynT2I-Eval: A Dynamic Evaluation Framework for Text-to-Image Models (NeurIPS 2026 · similarity 0.88)
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.
