UniSpace: Unified Visual Representation and Scalable Multimodal Modeling

• URL: http://arxiv.org/abs/2608.08676v1
• Date: Sun, 09 Aug 2026 12:46:56 GMT
• Title: UniSpace: Unified Visual Representation and Scalable Multimodal Modeling
• Abstract summary: We study whether understanding, generation, and editing can be modeled in a single visual representation space built from a pretrained semantic ViT.<n>We introduce emphPatch Re parameterization, which preserves the original semantic pathway while adding a reconstruction-aware patch embedding.<n>The resulting unified representation preserves multimodal understanding while enabling high-fidelity image reconstruction and a favorable reconstruction--generation trade-off.
• Score: 38.3945115074626
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Semantic vision encoders have become a central visual interface for multimodal understanding and semantic conditioning in image generation. However, their final tokens discard fine-grained visual details, leading to poor pixel reconstruction and limiting their use in reconstruction-sensitive tasks such as image generation and editing. In this work, we ask whether understanding, generation, and editing can be modeled in a single visual representation space built from a pretrained semantic ViT. We show that the frozen Transformer blocks of a semantic ViT are not intrinsically unable to preserve visual details. Instead, the original patch parameterization drives the representation toward semantic abstraction, making fine-grained information difficult to recover from the final tokens. Based on this observation, we introduce \emph{Patch Reparameterization}, which preserves the original semantic pathway while adding a reconstruction-aware patch embedding that provides fine-grained visual information to the same frozen ViT blocks. The resulting unified representation preserves multimodal understanding while enabling high-fidelity image reconstruction and a favorable reconstruction--generation trade-off. We further scale this representation into \emph{UniSpace}, an 8B Mixture-of-Transformer-Experts model that performs understanding, generation, and editing in the same visual space without a separate VAE pathway. System-level evaluations demonstrate practical text-to-image generation and instruction-based image editing, showing that a reparameterized pretrained ViT can serve as a unified visual interface for scalable multimodal modeling.

Related papers

• VoT: Vision-of-Thought for Unified Multimodal Representation Alignment [38.89764677562606]
  Vision-of-Thought (VoT) is a framework that introduces a discrete visual-thinking layer between vision-language models (VLMs) and diffusion transformers (DiTs)<n>VoT generates discrete VoT tokens representing high-level visual plans, such as objects and layouts, before rendering pixels.<n>We train a specialized VoT tokenizer in the VLM semantic space with a closed-loop objective that combines VLM alignment, feature reconstruction, and vector-quantization losses.
  arXiv  Detail & Related papers  (2026-09-07T17:53:13Z)
• Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment [20.811984469709508]
  Aura is a unified framework for high-fidelity and identity-consistent video generation.<n>We introduce AI director-level captions that provide dense and structured descriptions of video content.<n>We build a high-quality video-subject image dataset through a dedicated data construction pipeline.
  arXiv  Detail & Related papers  (2026-07-05T13:54:33Z)
• Imagine Before You Draw: Visual Prompt Engineering for Image Generation [67.81347924426714]
  We propose Visual Prompt Engineering (VPE), which can be seamlessly integrated into internal frameworks.<n>We validate VPE across class-conditional generation, text-to-image generation, and image editing.<n>Results show that VPE can accelerate convergence, raise quality ceilings, and through internal integration, achieve substantially better editing preservation.
  arXiv  Detail & Related papers  (2026-06-03T05:01:36Z)
• UniCustom: Unified Visual Conditioning for Multi-Reference Image Generation [65.53694602893042]
  We propose a unified visual conditioning framework that fuses ViT and VAE features before VLM encoding.<n>Experiments on two multi-reference generation benchmarks demonstrate that UniCustom consistently improves subject consistency, instruction following, and compositional fidelity.
  arXiv  Detail & Related papers  (2026-05-12T13:10:05Z)
• Self-Corrected Image Generation with Explainable Latent Rewards [55.29175717238288]
  We propose xLARD, a self-correcting framework that guides generation through Explainable LAtent RewarDs.<n>xLARD introduces a lightweight corrector that refines latent representations based on structured feedback from model-generated references.<n> Experiments show that xLARD improves semantic alignment and visual fidelity while maintaining generative priors.
  arXiv  Detail & Related papers  (2026-03-26T02:59:35Z)
• Splicing ViT Features for Semantic Appearance Transfer [10.295754142142686]
  We present a method for semantically transferring the visual appearance of one natural image to another. Specifically, our goal is to generate an image in which objects in a source structure image are "painted" with the visual appearance of their semantically related objects in a target appearance image.
  arXiv  Detail & Related papers  (2022-01-02T22:00:34Z)
• Intriguing Properties of Vision Transformers [114.28522466830374]
  Vision transformers (ViT) have demonstrated impressive performance across various machine vision problems. We systematically study this question via an extensive set of experiments and comparisons with a high-performing convolutional neural network (CNN) We show effective features of ViTs are due to flexible receptive and dynamic fields possible via the self-attention mechanism.
  arXiv  Detail & Related papers  (2021-05-21T17:59:18Z)
• Fine-grained Image-to-Image Transformation towards Visual Recognition [102.51124181873101]
  We aim at transforming an image with a fine-grained category to synthesize new images that preserve the identity of the input image. We adopt a model based on generative adversarial networks to disentangle the identity related and unrelated factors of an image. Experiments on the CompCars and Multi-PIE datasets demonstrate that our model preserves the identity of the generated images much better than the state-of-the-art image-to-image transformation models.
  arXiv  Detail & Related papers  (2020-01-12T05:26:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.