Research map / Image generation / Text-to-image diffusion
Semantic Browsing: Controllable Diversity for Image Generation
Sara Dorfman, Maya Vishnevsky, Omer Dahary, Or Patashnik, Danny Cohen-Or
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 |
|---|---|---|---|
| induce structured semantic diversity via agentic JSON-prompt tree instead of stochastic VLM seeding | FIBO · MS-COCO 50 prompts | Vendi score: 2.60 -> 3.34 | Improved |
| same change | FIBO · MS-COCO 50 prompts | DINO similarity: 0.76 -> 0.61 | Improved |
| same change | FIBO · MS-COCO 50 prompts | VQAScore: 0.93 -> 0.90 | Hurt |
| same change vs CADS | FIBO · MS-COCO 50 prompts | Aesthetic score: 6.30 -> 6.52 | Improved |
| add Context Analyst agent (ablation) | Semantic Browsing · Hierarchical Consistency | consistency: 0.82 -> 0.87 | Improved |
| add Context Analyst agent (ablation) | Semantic Browsing · MS-COCO | VQAScore: 0.90 -> 0.90 | No clear effect |
| add Critic agent (ablation) | Semantic Browsing · MS-COCO | VQAScore: 0.87 -> 0.90 | Improved |
| add Critic agent (ablation) | Semantic Browsing · Hierarchical Consistency | consistency: 0.87 -> 0.87 | No clear effect |
| separate Brainstormer and Decision Maker roles (ablation) | Semantic Browsing · tree DINO distance (mean) | DINO distance: 0.362 -> 0.389 | Improved |
| swap Gemini 2.5 Flash VLM for ChatGPT-5.5 (ablation) | Semantic Browsing · MS-COCO | Vendi score: 3.34 -> 3.30 | No clear effect |
| increase tree depth (BF=1) from gallery 5 to 20 (ablation) | Semantic Browsing · scaling ablation | VQAScore: 0.81 -> 0.75 | Hurt |
| increase branching factor (D=1) from 5 to 20 (ablation) | Semantic Browsing · scaling ablation | Vendi score: 2.05 -> 3.13 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CADS — CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling (ICLR 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
- Generative Powers of Ten (CVPR 2024 · similarity 0.88)
- Elucidating the design space of language models for image generation (ICML 2025 · similarity 0.88)
- Localizing and Editing Knowledge In Text-to-Image Generative Models (ICLR 2024 · similarity 0.88)
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.87)
- SSMG: Spatial-Semantic Map Guided Diffusion Model for Free-Form Layout-to-Image Generation (AAAI 2024 · similarity 0.87)
- VersaGen: Unleashing Versatile Visual Control for Text-to-Image Synthesis (AAAI 2025 · similarity 0.87)
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.
