Research map / Image generation / Text-to-image diffusion
Textual-Visual Logic Challenge: Understanding and Reasoning in Text-to-Image Generation
Peixi Xiong, Michael A Kozuch, Nilesh Jain
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 Negative Pair Discriminator loss (Focal Informative) (ablation) | proposed T2I model · TV-Logic | RSIM: 69.91 -> 72.78 | Improved |
| add learned relation attention (Focal Relation) (ablation) | proposed T2I model · TV-Logic | RSIM: 66.12 -> 72.78 | Improved |
| add Cross-Modality Tower instead of direct concatenation (ablation) | proposed T2I model · TV-Logic | RSIM: 68.90 -> 72.78 | Improved |
| add learned relation attention (Focal Relation) (ablation) | proposed T2I model · TV-Logic | FID: 39.41 -> 40.60 | Hurt |
| proposed model versus strongest baseline LatteGAN | TV-Logic · TV-Logic | RSIM: 63.03 -> 72.78 | Improved |
| numerical representation prompts | all models · TV-Logic Numerical Rep | F1: Ours 62.12 vs Detail category 86.79 | Hurt |
| scenarios with multiple same-attribute entities or target identification | all models · TV-Logic ambiguity categories | RSIM: Ours 81.01 (Detail) -> 58.04 (Relational Ambiguity) | Hurt |
Closest papers by meaning
- MMMG: A Massive, Multidisciplinary, Multi-Tier Generation Benchmark for Text-to-Image Reasoning (NeurIPS 2025 · similarity 0.89)
- GIR-Bench: Versatile Benchmark for Generating Images with Reasoning (ICLR 2026 · similarity 0.88)
- Image Content Generation with Causal Reasoning (AAAI 2024 · similarity 0.88)
- Vision as a Dialect: Unifying Visual Understanding and Generation via Text-Aligned Representations (NeurIPS 2025 · similarity 0.88)
- LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Prompts (ICLR 2024 · similarity 0.88)
- The Narrow Gate: Localized Image-Text Communication in Native Multimodal Models (NeurIPS 2025 · 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.
