CVAICLNov 20, 2025

Thinking-while-Generating: Interleaving Textual Reasoning throughout Visual Generation

arXiv:2511.16671v114 citationsh-index: 18Has Code
Originality Highly original
AI Analysis

This addresses the lack of on-the-fly multimodal interaction in visual generation, offering a novel approach for AI researchers, though it is a preliminary study.

The paper tackles the problem of integrating textual reasoning into visual generation by introducing Thinking-while-Generating (TwiG), a framework that interleaves reasoning throughout the process, resulting in more context-aware and semantically rich outputs.

Recent advances in visual generation have increasingly explored the integration of reasoning capabilities. They incorporate textual reasoning, i.e., think, either before (as pre-planning) or after (as post-refinement) the generation process, yet they lack on-the-fly multimodal interaction during the generation itself. In this preliminary study, we introduce Thinking-while-Generating (TwiG), the first interleaved framework that enables co-evolving textual reasoning throughout the visual generation process. As visual content is progressively generating, textual reasoning is interleaved to both guide upcoming local regions and reflect on previously synthesized ones. This dynamic interplay produces more context-aware and semantically rich visual outputs. To unveil the potential of this framework, we investigate three candidate strategies, zero-shot prompting, supervised fine-tuning (SFT) on our curated TwiG-50K dataset, and reinforcement learning (RL) via a customized TwiG-GRPO strategy, each offering unique insights into the dynamics of interleaved reasoning. We hope this work inspires further research into interleaving textual reasoning for enhanced visual generation. Code will be released at: https://github.com/ZiyuGuo99/Thinking-while-Generating.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes