Self-Correcting Coupled Markov Jump Processes Enable Joint Image-Text Generation and Understanding

Jul 16, 2026

A new framework, Self-Correcting Coupled Markov Jump Processes (SC-CMJP), is introduced to enable concurrent image and text generation by coupling masked diffusion models across modalities. The associated training-free sampler, CO₂Jump, demonstrates state-of-the-art performance on joint multimodal tasks such as image editing and visual reasoning, as shown on newly released large-scale benchmarks.

Why it matters: This work represents a significant advance in multimodal AI by enabling real-time, cross-modal correction and coherent joint outputs, addressing limitations of previous systems that treated modalities separately.

Full story at: arXiv Machine Learning ↗