Bridging Modal Isolation in Interleaved Thinking: Supervising Modality Transitions via Stepwise Reinforcement

2026-06-11 · Tingyu Li, Le Zhou, Siyuan Li, Yujun Wu, Xinglong Xu, Jingxuan Wei, Conghui He, Cheng Tan arxiv

Interleaved thinking, where a unified multimodal model alternates between textual reasoning and visual generation, has shown promise on spatial and physical tasks. However, in complex long-chain scenarios, we identify a fundamental failure mode: generated images diverge from the textual context while subsequent text ignores the visual evidence, causing the two modalities to alternate without genuinely informing each other. We term this Modal Isolation and attribute it to compounding information loss at modality boundaries. We decompose each reasoning cycle into atomic operations and define modality transition loss, quantifying cross-modal hallucination (text-to-image) and visual utilization deficit (image-to-text) at each boundary. We propose MoTiF (Modality Tiransition Fidelity), a two-stage training framework that directly optimizes these transitions: Reflective SFT trains the model to detect and recover from erroneous visual outputs; Flow-GRPO improves image generation fidelity via reinforcement learning. All training signals in MoTiF derive from transition-level fidelity rather than end-task accuracy. Across four visual puzzle benchmarks, this transition-level supervision substantially improves both cross-modal coherence and final task accuracy. The results demonstrate that effective interleaved reasoning requires explicit structural supervision at modality boundaries, not merely scaling or end-task optimization.

📄 PDF Abstract BibTeX arXiv:2606.12886

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning Image Generation

Bridging Interleaved Multi-Modal Reasoning as a Unified Decision Process

2026-07-04 · Zican Hu, Xuyang Hu, Yiming Liu, Zuwei Long 외 hf

Unified multi-modal models (UMMs) have shown promising interleaved text-image reasoning capabilities, yet effectively optimizing such multi-turn generation via reinforcement learning (RL) remains an open challenge. Exist…

Reinforcement Learning Spatial Reasoning Image Generation Image Denoising

OPLD: On-Policy Latent Distillation for Multimodal Reasoning

2026-07-30 · Shoutai Zhu, Tianyang Xu, Bin Sun, Mingyuan Xu 외 arxiv

Interleaved multimodal Chain-of-Thought (CoT) improves visual reasoning by incorporating auxiliary visual evidence into intermediate reasoning. However, existing approaches remain constrained by externally defined reason…

Multimodal Reasoning Visual Reasoning

Reliable Thinking with Images

2026-02-13 · Haobin Li, Yutong Yang, Yijie Lin, Xiang Dai 외 arxiv

As a multimodal extension of Chain-of-Thought (CoT), Thinking with Images (TWI) has recently emerged as a promising avenue to enhance the reasoning capability of Multi-modal Large Language Models (MLLMs), which generates…

LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning

2026-05-21 · Yifan Dai, Zhenhua Wu, Bohan Zeng, Daili Hua 외 arxiv

Joint audio-visual reasoning is essential for omnimodal understanding, yet current multimodal large language models (MLLMs) still struggle when reasoning requires fine-grained evidence from both modalities. A central lim…

Visual Reasoning

ThinkingVLA: Interleaved Vision and Language Reasoning for Robotic Manipulation

2026-06-16 · Tianyi Lu, Hui Zhang, Zijie Diao, Junke Wang 외 arxiv

Most Vision-Language-Action (VLA) models map observations directly to actions without explicit reasoning, limiting their capacity for reasoning-intensive long-horizon tasks. To address this, existing approaches adopt Cha…