mlsanityData to 14 Oct 2026

KoVRE: Training an Efficient Embedding Model for Korean Visual Document Retrieval

arXiv 2608.01389 ↗First posted 2 Aug 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Visual Document Retrieval (VDR) directly matches text queries against document images, preserving visual and structural information that may be lost during text extraction. However, existing VDR models and training resources remain predominantly English-centric, while many high-performing systems rely on massive backbones or storage-intensive multi-vector representations. To address these limitations, we introduce KoVRE: Korean Visual Document Retrieval Embedding, a single-vector retriever for Korean visual documents, alongside a comprehensive training recipe. We train the model on 708,729 Korean and English query-page pairs using positive-aware hard-negative mining and conduct controlled analyses of training-data composition, hard-negative treatment, and reranker-based knowledge distillation. Across Korean visual document retrieval benchmarks, our 2B model substantially improves over the base backbone model, outperforming both its 8B single-vector counterpart and a strong multi-vector baseline. These results demonstrate that targeted bilingual supervision and our carefully designed training strategies can produce a highly effective Korean VDR model across diverse document domains, without requiring a scaled-up backbone or multi-vector representations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 33 of its 43 references that are in the map. First posted on arXiv 2 Aug 2026.

Builds on

33 of 43 references.

References · 33

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2020Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrievalcited by 849
  3. 2021ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interactioncited by 476
  4. 2020RocketQA: An Optimized Training Approach to Dense Passage Retrieval for Open-Domain Question Answeringcited by 413
  5. 2021TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Financecited by 408
  6. 2024VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documentscited by 245
  7. 2026Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and Rankingcited by 230
  8. 2021RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-rankingcited by 176
  9. 2020Improving Efficient Neural Ranking Models with Cross-Architecture Knowledge Distillationcited by 145
  10. 2025EmbeddingGemma: Powerful and Lightweight Text Representationscited by 139
  11. 2021Scaling Deep Contrastive Learning Batch Size under Memory Limited Setupcited by 127
  12. 2025MMTEB: Massive Multilingual Text Embedding Benchmarkcited by 125
21 more
  1. 2024Mitigating Catastrophic Forgetting in Large Language Models with Self-Synthesized Rehearsalcited by 118
  2. 2024M3DocRAG: Multi-modal Retrieval is What You Need for Multi-page Multi-document Understandingcited by 111
  3. 2024Unifying Multimodal Retrieval via Document Screenshot Embeddingcited by 110
  4. 2024NV-Retriever: Improving text embedding models with effective hard-negative miningcited by 84
  5. 2025jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrievalcited by 76
  6. 2025ViDoRAG: Visual Document Retrieval-Augmented Generation via Dynamic Iterative Reasoning Agentscited by 66
  7. 2025ViDoRe Benchmark V2: Raising the Bar for Visual Retrievalcited by 61
  8. 2025VDocRAG: Retrieval-Augmented Generation over Visually-Rich Documentscited by 59
  9. 2025LLaVE: Large Language and Vision Embedding Models with Hardness-Weighted Contrastive Learningcited by 55
  10. 2024GISTEmbed: Guided In-sample Selection of Training Negatives for Text Embedding Fine-tuningcited by 40
  11. 2024jina-clip-v2: Multilingual Multimodal Embeddings for Text and Imagescited by 37
  12. 2025REAL-MM-RAG: A Real-World Multi-Modal Retrieval Benchmarkcited by 35
  13. 2025Llama Nemoretriever Colembed: Top-Performing Text-Image Retrieval Modelcited by 27
  14. 2025Towards Storage-Efficient Visual Document Retrieval: An Empirical Study on Reducing Patch-Level Embeddingscited by 26
  15. 2026ViDoRe V3: A Comprehensive Evaluation of Retrieval Augmented Generation in Complex Real-World Scenarioscited by 26
  16. 2025ModernVBERT: Towards Smaller Visual Document Retrieverscited by 21
  17. 2025UNIDOC-BENCH: A Unified Benchmark for Document-Centric Multimodal RAGcited by 21
  18. 2026Nemotron ColEmbed V2: Top-Performing Late Interaction Embedding Models for Visual Document Retrievalcited by 10
  19. 2025MIRACL-VISION: A Large, multilingual, visual document retrieval benchmarkcited by 9
  20. 2026jina-embeddings-v5-omni: Geometry-preserving Embeddings via Locked Aligned Towerscited by 5
  21. 2025SDS KoPub VDR: A Benchmark Dataset for Visual Document Retrieval in Korean Public Documentscited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025DocPruner: A Storage-Efficient Framework for Multi-Vector Visual Document Retrieval via Adaptive Patch-Level Embedding Pruning
  2. 2026EVIE: Evidence-Vector-Informed Embeddings for Visual Document Retrieval ↗
  3. 2026MURE: Hierarchical Multi-Resolution Encoding via Vision-Language Models for Visual Document Retrieval
  4. 2026CausalEmbed: Auto-Regressive Multi-Vector Generation in Latent Space for Visual Document Embedding
  5. 2024ColPali: Efficient Document Retrieval with Vision Language Models
  6. 2026AnchorFold: A Focus-Then-Fold Framework via Recursive Attention Propagation for Efficient Multi-Vector Visual Document Retrieval
  7. 2025SERVAL: Surprisingly Effective Zero-Shot Visual Document Retrieval Powered by Large Vision and Language Models
  8. 2026Unveil: Unified Visual-Textual Integration and Distillation for Multi-modal Document Retrieval