Research map / Image generation / Text-to-image diffusion
OctoT2I: A Self-Evolving Agentic Text-to-Image Router
Xu Jiang, Bin Chen, Gehui Li, Yule Duan, Ronggang Wang, Jian Zhang
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 self-evolved knowledge instead of GPT internal knowledge (ablation) | OctoT2I · GenEval | overall score: 0.85 -> 0.96 | Improved |
| add self-evolved knowledge instead of hand-crafted prior (ablation) | OctoT2I · GenEval | overall score: 0.93 -> 0.96 | Improved |
| add knowledge-driven decision policy instead of random tool selection (ablation) | OctoT2I · T2I-CompBench++ | average score: 0.5379 -> 0.6618 | Improved |
| add exploration space pruning (ablation) | OctoT2I · GenEval | explored prompts: 1270 -> 370 | Improved |
| add exploration space pruning (ablation) | OctoT2I · GenEval | exploration time (s): 6857.4 -> 2328.7 | Improved |
| add exploration space pruning (ablation) | OctoT2I · GenEval | overall score: 0.96 -> 0.96 | No clear effect |
| raise quality threshold theta (ablation) | OctoT2I · GenEval | performance and latency: most significant gain up to theta=0.8 | Mixed |
| add one new tool (FLUX.1-dev) with 60 exploration prompts | OctoT2I · WISE | average score: 0.54 -> 0.61 | Improved |
| add second new tool (gpt-image-1) | OctoT2I · WISE | average score: 0.61 -> 0.71 | Improved |
| swap controller LLM to GPT-5 | OctoT2I · WISE | average score: 0.54 -> 0.60 | Improved |
| swap controller LLM to Qwen3-VL-235B-A22B | OctoT2I · WISE | average score: 0.54 -> 0.55 | No clear effect |
| use OctoT2I instead of Flow-GRPO | OctoT2I · GenEval (efficiency) | average time (s): 19.07 -> 10.02 | 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.
- GenArtist — GenArtist: Multimodal LLM as an Agent for Unified Image Generation and Editing (NeurIPS 2024) · better on 3 of 3 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
- T2I-Adapter: Learning Adapters to Dig Out More Controllable Ability for Text-to-Image Diffusion Models (AAAI 2024 · similarity 0.87)
- Generation Navigator: A State-Aware Agentic Framework for Image Generation (NeurIPS 2026 · similarity 0.87)
- KOALA: Empirical Lessons Toward Memory-Efficient and Fast Diffusion Models for Text-to-Image Synthesis (NeurIPS 2024 · similarity 0.87)
- Gen-Searcher: Reinforcing Agentic Search for Image Generation (NeurIPS 2026 · similarity 0.86)
- Arena-T2I Hard: Benchmarking and Improving Faithfulness with Dependency-Aware Checklist Rewards (NeurIPS 2026 · similarity 0.86)
- DynT2I-Eval: A Dynamic Evaluation Framework for Text-to-Image Models (NeurIPS 2026 · 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.
