Latent Visual States for Efficient Multimodal Reasoning

• URL: http://arxiv.org/abs/2606.24233v1
• Date: Tue, 23 Jun 2026 07:22:22 GMT
• Title: Latent Visual States for Efficient Multimodal Reasoning
• Abstract summary: EVA (LatEnt Visual StAtes) is a novel framework that generates continuous latent visual representations.<n>These internal representations manifest as an adaptive sequence of Latent_slot tokens.<n> EVA achieves significant performance gains while enhancing inference efficiency.
• Score: 58.350643620457
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: The integration of visual evidence has significantly enhanced the capabilities of large multimodal models. However, this integration predominantly relies on generating discrete outputs (etc., code or box coordinates) to invoke external tools, a process that introduces rigid dependencies and substantial latency. To overcome these limitations, we propose {EVA} (LatEnt Visual StAtes), a novel framework that natively generates continuous latent visual representations. These internal representations manifest as an adaptive sequence of Latent\_slot tokens, serving as intermediate visual thoughts during the reasoning process. These Latent\_slot tokens are then trained end-to-end with the discrete text tokens. This co-optimization, notably, causes extreme policy deviation in the 'transition window' following the Latent\_slot tokens. We develop D-GSPO (Decouple-GSPO) to target this root cause by decoupling the optimization of latent and discrete components. To support SFT, we construct EVA-230K, a high-quality text-image interleaved CoT dataset encompassing a diverse range of real-world scenes, documents, charts and OCR tasks. Extensive experiments across multiple benchmarks confirm that EVA achieves significant performance gains while enhancing inference efficiency.

Related papers

• Beyond Visual Cues: Semantic-Driven Token Filtering and Expert Routing for Anytime Person ReID [10.58405139566836]
  Any-Time Person Re-identification (AT-ReID) necessitates the robust retrieval of target individuals under arbitrary conditions.<n>Existing methods are highly relying on pure visual features, which are prone to change due to environmental and time factors.<n>We propose a novel framework that leverages the ability of Large Vision-Language Models (LVLMs) to generate identity consistency text.
  arXiv  Detail & Related papers  (2026-04-16T14:49:30Z)
• MURE: Hierarchical Multi-Resolution Encoding via Vision-Language Models for Visual Document Retrieval [111.99106496142474]
  Visual Document Retrieval (VDR) requires representations that capture both fine-grained visual details and global document structure.<n>Existing VDR models struggle to balance effectiveness and efficiency when processing high-resolution documents.<n>We propose a new X-VisEmb paradigm that progresses from multi-resolution sampling and encoding, through cross-granularity feature fusion, to adaptive representation distillation.
  arXiv  Detail & Related papers  (2026-03-07T15:17:22Z)
• Perception-Aware Multimodal Spatial Reasoning from Monocular Images [57.42071289037214]
  spatial reasoning from monocular images is essential for autonomous driving.<n>Current Vision-Language Models (VLMs) still struggle with fine-grained geometric perception.<n>We propose a perception-aware multimodal reasoning framework that equipsVLMs with explicit object-centric grounding ability.
  arXiv  Detail & Related papers  (2026-03-07T02:05:12Z)
• CausalEmbed: Auto-Regressive Multi-Vector Generation in Latent Space for Visual Document Embedding [71.88471147281406]
  We propose an auto-regressive generation approach, CausalEmbed, for constructing multi-vector embeddings.<n>By incorporating iterative margin loss during contrastive training, CausalEmbed encourages embedding models to learn compact and well-structured representations.<n>Our method enables efficient VDR tasks using only dozens of visual tokens, achieving a 30-155x reduction in token count.
  arXiv  Detail & Related papers  (2026-01-29T04:47:27Z)
• End-to-End Vision Tokenizer Tuning [73.3065542220568]
  The vision tokenizer optimized for low-level reconstruction is to downstream tasks requiring varied representations and semantics.<n>The loss of the vision tokenization can be the representation bottleneck for target tasks.<n>We propose ETT, an end-to-end vision tokenizer tuning approach that enables joint optimization between vision tokenization and target autoregressive tasks.
  arXiv  Detail & Related papers  (2025-05-15T17:59:39Z)
• ContextFormer: Redefining Efficiency in Semantic Segmentation [48.81126061219231]
  Convolutional methods, although capturing local dependencies well, struggle with long-range relationships.<n>Vision Transformers (ViTs) excel in global context capture but are hindered by high computational demands.<n>We propose ContextFormer, a hybrid framework leveraging the strengths of CNNs and ViTs in the bottleneck to balance efficiency, accuracy, and robustness for real-time semantic segmentation.
  arXiv  Detail & Related papers  (2025-01-31T16:11:04Z)
• Vision-Enhanced Semantic Entity Recognition in Document Images via Visually-Asymmetric Consistency Learning [19.28860833813788]
  Existing models commonly train a visual encoder with weak cross-modal supervision signals. We propose a novel textbfVisually-textbfAsymmetric cotextbfNsistentextbfCy textbfLearning (textscVancl) approach to capture fine-grained visual and layout features.
  arXiv  Detail & Related papers  (2023-10-23T10:37:22Z)
• Good Visual Guidance Makes A Better Extractor: Hierarchical Visual Prefix for Multimodal Entity and Relation Extraction [88.6585431949086]
  We propose a novel Hierarchical Visual Prefix fusion NeTwork (HVPNeT) for visual-enhanced entity and relation extraction. We regard visual representation as pluggable visual prefix to guide the textual representation for error insensitive forecasting decision. Experiments on three benchmark datasets demonstrate the effectiveness of our method, and achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2022-05-07T02:10:55Z)

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.