CyCLeGen: Cycle-Consistent Layout Prediction and Image Generation in Vision Foundation Models
This addresses the challenge of separate perception and synthesis modules in vision models, offering a more efficient and introspective approach for vision-language tasks.
The paper tackles the problem of integrating image understanding and generation in vision-language models by proposing CyCLeGen, a unified autoregressive framework that uses cycle-consistent learning, resulting in significant gains across diverse benchmarks.
We present CyCLeGen, a unified vision-language foundation model capable of both image understanding and image generation within a single autoregressive framework. Unlike existing vision models that depend on separate modules for perception and synthesis, CyCLeGen adopts a fully integrated architecture that enforces cycle-consistent learning through image->layout->image and layout->image->layout generation loops. This unified formulation introduces two key advantages: introspection, enabling the model to reason about its own generations, and data efficiency, allowing self-improvement via synthetic supervision under a reinforcement learning objective guided by cycle consistency. Extensive experiments show that CyCLeGen achieves significant gains across diverse image understanding and generation benchmarks, highlighting the potential of unified vision-language foundation models.