STBridge: Shared-Target Alignment for Bridging Understanding and Generation in UMMs
Computer Vision and Pattern Recognition
Summary
The authors looked at models that both understand and create images, called unified multimodal models (UMMs). They found these models often describe an image correctly but then make edits that don't match the description well. To fix this, the authors created STBridge, a method that helps the model keep its description and image changes closely connected by using a shared target goal. Their tests showed STBridge makes the models better at matching what they say with what they create, especially for detailed image parts. This work suggests that sharing a clear goal between understanding and generating helps improve image editing in these models.
Authors
Ye Wang, Hongjun Wang, Hao Fang, Tongyuan Bai, Zuwei Long, Peixian Chen, Wei Liu, Weibo Gu, Xing Sun, Rui Ma
Abstract
Unified multimodal models (UMMs) aim to integrate visual understanding and generation within a single architecture, but architectural unification alone does not ensure semantic consistency. A model may describe the intended target correctly while generating an inconsistent edit. This exposes an understanding-generation alignment gap: linguistic and visual outputs live in different spaces, yet should be governed by the same target semantics. We study this gap in image editing, where an instruction defines a target state that can be both described and visually realized. Given a source image and an edit instruction, we compare a UMM's target caption with its edited image to test whether the two outputs converge on the same result. Our analysis shows that existing UMMs remain weakly aligned, especially for fine-grained entities, attributes, spatial relations, and local details, indicating that semantic unification is not achieved by architecture alone. To bridge this gap, we propose STBridge, a shared-target alignment framework that connects understanding and generation through a common target state. Here the target caption expresses the desired visual result, while the edited image realizes it visually, replacing separate task-specific paths with a shared information flow from target expression to target realization. STBridge follows an align-then-optimize strategy: supervised fine-tuning first establishes the shared-target channel, and sequential reinforcement learning further refines target-centered coordination. Across visual understanding, image generation, and image editing benchmarks, STBridge consistently improves over the initialization model. Alignment analysis confirms that STBridge narrows the gap between what the model describes and what it generates, demonstrating shared-target alignment as an effective post-training strategy for bridging understanding and generation in UMMs.