{"total":39,"items":[{"citing_arxiv_id":"2607.00596","ref_index":2,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Semantic-Guided Reading Order Reconstruction in Historical Armenian Newspapers with LLMs","primary_cat":"cs.CV","submitted_at":"2026-07-01T08:19:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Hybrid semantic-LLM method for reading order reconstruction in Armenian historical newspapers outperforms baselines on a new 66-page dataset while releasing a specialized Tesseract OCR model.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.27047","ref_index":5,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"NuclearQAv2: A Structured Benchmark for Evaluating Domain-Science Competence in Large Language Models","primary_cat":"cs.CL","submitted_at":"2026-06-25T13:52:16+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"NuclearQAv2 is a hybrid-constructed benchmark dataset for evaluating LLM competence in nuclear engineering knowledge using three question types.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.25363","ref_index":9,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"TheoremGraph: Bridging Formal and Informal Mathematics","primary_cat":"cs.IR","submitted_at":"2026-06-24T03:47:05+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"TheoremGraph builds a unified statement-level dependency graph across informal arXiv math and formal Lean code via parsing, embeddings, and LLM validation, releasing the data and APIs for search and retrieval.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.25343","ref_index":10,"ref_count":2,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Invoice Haystack: Benchmarking Document Retrieval and Visual Question Answering Under Strong Visual Homogeneity","primary_cat":"cs.CV","submitted_at":"2026-06-24T03:17:30+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Presents Invoice Haystack benchmark for homogeneous document retrieval and VL-RAG hybrid framework achieving 60% Recall@1 and up to 13.5 point gains over prior methods.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.24734","ref_index":188,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Task Decomposition for Efficient Annotation","primary_cat":"cs.CL","submitted_at":"2026-06-23T15:58:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"Decomposing annotation tasks using centers from centering theory reduces aggregate inferential load via a degrees-of-freedom model and enables better sub-task allocation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.23344","ref_index":18,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"RT-DocLayout: Real-Time End-to-End Document Layout Analysis with Reading Order in the Wild","primary_cat":"cs.CV","submitted_at":"2026-06-22T13:48:39+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Presents RT-DocLayout, a 33M-parameter end-to-end model extending RT-DETR that unifies layout classification, detection, segmentation, and reading-order prediction at 132.1 FPS with claimed SOTA results on public benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.23050","ref_index":10,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Unlimited OCR Works","primary_cat":"cs.CV","submitted_at":"2026-06-22T09:01:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"The paper proposes Unlimited OCR using Reference Sliding Window Attention (R-SWA) to achieve constant KV cache for efficient transcription of long multi-page documents.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02915","ref_index":5,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Any2Poster: Any-Source Poster Generation Across Modalities and Domains","primary_cat":"cs.CV","submitted_at":"2026-06-01T21:41:21+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Any2Poster Bench tests poster generation from 8 modalities and 5 domains using quizzes and VLM judgments; Any2Poster Agent reaches 87% accuracy and beats prior paper-only methods.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.00902","ref_index":22,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Ryze: Evidence-Enriched Data Synthesis from Biomedical Papers","primary_cat":"cs.AI","submitted_at":"2026-05-30T21:54:02+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Ryze automates evidence-enriched QA synthesis from biomedical papers to produce BioVLM-8B, which reaches 48.0% weighted accuracy on LAB-Bench (+12.6pp over base, +3.8pp over GPT-5.2) at under $200 cost.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.00224","ref_index":17,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"An AI-ready, Polarized Electron-Positron Collision Dataset","primary_cat":"hep-ex","submitted_at":"2026-05-29T18:00:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Release of an AI-ready dataset containing approximately 660,000 reconstructed polarized e+e- collision events at 91.2 GeV from the SLD experiment, translated from legacy formats with accompanying digitized documentation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.22100","ref_index":15,"ref_count":2,"confidence":0.98,"is_internal_anchor":true,"paper_title":"MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsing","primary_cat":"cs.AI","submitted_at":"2026-05-21T07:36:41+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MPDocBench-Parse provides 433 annotated multi-page documents and an evaluation protocol covering text/table/formula extraction, merging, figure extraction, reading order, and heading hierarchy for realistic document parsing.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.21611","ref_index":2,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"UniVL: Unified Vision-Language Embedding for Spatially Grounded Contextual Image Generation","primary_cat":"cs.CV","submitted_at":"2026-05-20T18:17:50+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"UniVL unifies vision and language into one mask-rendered input processed by an OCR backbone to condition diffusion models for spatially grounded image generation without a standalone text encoder.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.19866","ref_index":3,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Structured Layout Priors for Robust Out-of-Distribution Visual Document Understanding","primary_cat":"cs.CV","submitted_at":"2026-05-19T13:58:24+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Injecting pre-computed layout priors from RT-DETR into VLM prompts raises markdown F1 from 0.37 to 0.92 on a 10k-page OOD benchmark and cuts infinite-loop failures across domains.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.12623","ref_index":11,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DocAtlas: Multilingual Document Understanding Across 80+ Languages","primary_cat":"cs.CL","submitted_at":"2026-05-12T18:09:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DocAtlas introduces model-free rendering pipelines to create DocTag-annotated datasets across 82 languages and shows DPO adaptation improves multilingual performance without base-language degradation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.18818","ref_index":3,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Operationalizing Document AI: A Microservice Architecture for OCR and LLM Pipelines in Production","primary_cat":"cs.AI","submitted_at":"2026-05-12T13:07:34+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"Describes a microservice architecture for production document AI pipelines with OCR and LLMs, reporting that OCR dominates latency and GPU inference capacity limits concurrency.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.10341","ref_index":20,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"PaperFit: Vision-in-the-Loop Typesetting Optimization for Scientific Documents","primary_cat":"cs.AI","submitted_at":"2026-05-11T10:43:41+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"PaperFit uses rendered page images in a closed loop to diagnose and repair typesetting defects in LaTeX documents, outperforming baselines on a new benchmark of 200 papers.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"stage in the document automation pipeline. Date:May 12, 2026 Correspondence:Cheng Tan,tancheng@pjlab.org.cn Code Dataset 1 Introduction The past decade has witnessed remarkable progress in document automation. Format conversion tools such as Pandoc [139] enable structural transformation from Word and Markdown to LATEX. Document understanding models [20, 190, 49] can reconstruct LATEX source code from PDF files. Recent large language models (LLMs) can generate complete LATEX document frameworks directly from natural descriptions [163, 218]. We refer to this stage collectively asstructural formatting, whose primary objective is to produce compilable .tex files. However, compilation success does not guarantee visual"},{"citing_arxiv_id":"2604.23813","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"ShredBench: Evaluating the Semantic Reasoning Capabilities of Multimodal LLMs in Document Reconstruction","primary_cat":"cs.CV","submitted_at":"2026-04-26T17:26:06+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"ShredBench shows state-of-the-art MLLMs perform well on intact documents but suffer sharp drops in restoration accuracy as fragmentation increases to 8-16 pieces, indicating insufficient cross-modal semantic reasoning for VRDU.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"can leverage internal language priors to maintain coherence across visual discontinuities. To systematically evaluate this, we propose SHREDBENCH, a benchmark characterized by three key dimensions:(1) Multi-Granularity Com- plexity.We partition images into 8, 12, and 16 fragments. This hierarchy enables the analysis of how visual entropy correlates with performance degradation.(2) Diverse Scenarios.Comprising 756 documents, our dataset spans English and Chi- nese text, source code (strict syntax), and tables (complex 2D structure). Tables and code are no- tably difficult, requiring models to restore rigid indentation and alignment-a challenge even for specialized models (Zhang et al., 2024).(3) Exten- sive Experiments.We evaluate state-of-the-art pro-"},{"citing_arxiv_id":"2604.18584","ref_index":33,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"MathNet: a Global Multimodal Benchmark for Mathematical Reasoning and Retrieval","primary_cat":"cs.AI","submitted_at":"2026-04-20T17:59:49+00:00","verdict":"ACCEPT","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"MathNet delivers the largest multilingual Olympiad math dataset and benchmarks where models like Gemini-3.1-Pro reach 78% on solving but embedding models struggle on equivalent problem retrieval, with retrieval augmentation yielding up to 12% gains.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.17680","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"MasterSet: A Large-Scale Benchmark for Must-Cite Citation Recommendation in the AI/ML Literature","primary_cat":"cs.IR","submitted_at":"2026-04-20T00:34:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"MasterSet is a new large-scale benchmark for must-cite citation recommendation in AI/ML, using LLM-annotated tiers on 150k papers and Recall@K evaluation.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.07530","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"The Shrinking Lifespan of LLMs in Science","primary_cat":"cs.DL","submitted_at":"2026-04-08T19:12:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"LLM adoption in science follows a compressing inverted-U trajectory where release year predicts time-to-peak and lifespan better than model attributes.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.06079","ref_index":9,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Scientific Graphics Program Synthesis via Dual Self-Consistency Reinforcement Learning","primary_cat":"cs.CV","submitted_at":"2026-04-07T16:58:14+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SciTikZer-8B uses a new dataset, benchmark, and dual self-consistency RL to generate TikZ code for scientific graphics, outperforming much larger models like Gemini-2.5-Pro.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"chart-to-code fidelity and scalability. However, imperative charting often abstracts geometric details. In contrast, TikZ isdeclarativeand compilation-critical, requiring explicit spatial specifications that challenge current MLLMs. Automated TikZ Generation.Image-to-markup generation has matured in text and math domains (e.g., Im2Latex-100K [14], Nougat [9]). However, TikZ recovery fundamentally differs from image vectorization [28], which yields unstructured primitives lacking the semantic topology essential for scientific diagrams. Direct TikZ synthesis remains underexplored due to data quality bottlenecks. Text- to-TikZ methods such as TikZilla [25] and AutomaTikZ [5] generate compilable code from language,"},{"citing_arxiv_id":"2604.04771","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"MinerU2.5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scale","primary_cat":"cs.CV","submitted_at":"2026-04-06T15:44:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A fixed 1.2B model trained via diversity-aware sampling, cross-model verification, annotation refinement, and progressive stages achieves new state-of-the-art document parsing accuracy of 95.69 on OmniDocBench v1.6.","context_count":1,"top_context_role":"baseline","top_context_polarity":"baseline","context_text":"Document parsing converts unstructured documents such as PDFs into structured, machine-readable formats (e.g. Markdown), serving as critical infrastructure for LLM training data pipelines [49, 29, 35] and retrieval-augmented generation systems [15, 37, 48]. As end-to-end approaches based on vision- language models (VLMs) progressively replace traditional pipeline systems [2, 39, 25], research has focused predominantly on architectural innovation and inference efficiency, leading to rapid score convergence among top models on standard benchmarks. Yet this convergence raises a deeper question: what constitutes the remaining performance bottleneck? Our cross-analysis of parsing results from multiple state-of-the-art models-spanning diverse architec-"},{"citing_arxiv_id":"2604.04324","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"RESCORE: LLM-Driven Simulation Recovery in Control Systems Research Papers","primary_cat":"cs.AI","submitted_at":"2026-04-06T00:13:14+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"RESCORE recovers task-coherent simulations from 40.7% of 500 CDC papers via a three-component LLM agent pipeline and claims a 10X speedup over manual human replication.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.03476","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"Fine-tuning DeepSeek-OCR-2 for Molecular Structure Recognition","primary_cat":"cs.CV","submitted_at":"2026-04-03T21:42:54+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"MolSeek-OCR reaches exact SMILES matching accuracy comparable to leading image-to-sequence OCSR models after two-stage fine-tuning on PubChem renderings and USPTO-MOL patent images, but remains below image-to-graph state-of-the-art.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.09617","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"AdaQE-CG: Adaptive Query Expansion for Web-Scale Generative AI Model and Data Card Generation","primary_cat":"cs.AI","submitted_at":"2026-03-16T04:02:56+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"AdaQE-CG uses context-aware adaptive query expansion and inter-card knowledge transfer from a MetaGAI Pool to generate higher-quality model and data cards than prior methods, validated on the new expert-annotated MetaGAI-Bench.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.22755","ref_index":24,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"RADIANT-LLM: an Agentic Retrieval Augmented Generation Framework for Reliable Decision Support in Safety-Critical Nuclear Engineering","primary_cat":"cs.IR","submitted_at":"2026-03-04T01:30:22+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"RADIANT-LLM is a local-first multi-modal RAG system with provenance tracking that delivers lower hallucination rates than general LLMs on nuclear engineering benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.06179","ref_index":22,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"ARIA: Adaptive Retrieval Intelligence Assistant -- A Multimodal RAG Framework for Domain-Specific Engineering Education","primary_cat":"cs.IR","submitted_at":"2026-02-04T01:08:24+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"ARIA is a multimodal RAG framework that filters domain-specific questions with 97.5% accuracy and outperforms ChatGPT-5 on pedagogical quality for a university civil engineering course.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2602.01785","ref_index":14,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"CodeOCR: On the Effectiveness of Vision Language Models in Code Understanding","primary_cat":"cs.CL","submitted_at":"2026-02-02T08:10:21+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Multimodal LLMs process code as images to achieve up to 8x token compression, with visual cues like syntax highlighting aiding tasks and clone detection remaining resilient or even improving under compression.","context_count":1,"top_context_role":"method","top_context_polarity":"use_method","context_text":"(2)Text, where code is provided as plain text tokens, representing the standard text-based approach. The NoCtx baseline is not applicable for Code Summarization and Clone Detection, as these tasks require source code to be summarized or compared. 3.5 Implementation Details We implement our experiments in Python using a custom rendering pipeline built on Pygments [14] for syntax tokenization and Pillow [25] for image generation and processing. The base images are rendered with the default monospaced font from Visual Studio Code [60], at a font size of 40 pixels as suggested by prior work [55, 107], line height of 1.0, and margin of 1% of the page width. For syntax highlighting, we adopt the \"Default Light\" theme from Visual Studio Code [60]."},{"citing_arxiv_id":"2511.22490","ref_index":5,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"SciPostGen: Bridging the Gap between Scientific Papers and Poster Layouts","primary_cat":"cs.CV","submitted_at":"2025-11-27T14:27:33+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SciPostGen supplies a paired dataset linking paper structure to poster layouts and shows that retrieval of matching layouts improves generation while respecting user constraints.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2510.18234","ref_index":6,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DeepSeek-OCR: Contexts Optical Compression","primary_cat":"cs.CV","submitted_at":"2025-10-21T02:41:44+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"DeepSeek-OCR compresses text contexts up to 20x via 2D optical mapping while achieving 97% OCR accuracy below 10x and 60% at 20x, outperforming prior OCR tools with fewer vision tokens.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"OCR, particularly document parsing task, has been a highly active topic in the image-to-text domain. With the advancement of VLMs, a large number of end-to-end OCR models have emerged, fundamentally transforming the traditional pipeline architecture (which required separate detection and recognition expert models) by simplifying OCR systems. Nougat [ 6] first employs end-to-end framework for academic paper OCR on arXiv, demonstrating the potential of models in handling dense perception tasks. GOT-OCR2.0 [38] expands the scope of OCR2.0 to include more synthetic image parsing tasks and designs an OCR model with performance-efficiency trade-offs, further highlighting the potential of end-to-end OCR re-"},{"citing_arxiv_id":"2509.22186","ref_index":4,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing","primary_cat":"cs.CV","submitted_at":"2025-09-26T10:45:48+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MinerU2.5 uses a two-stage decoupled vision-language architecture to achieve state-of-the-art document parsing accuracy with lower computational overhead than existing general and domain-specific models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2507.08458","ref_index":38,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"A document is worth a structured record: Principled inductive bias design for document recognition","primary_cat":"cs.CV","submitted_at":"2025-07-11T10:02:08+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"Introduces a method to design structure-specific relational inductive biases for a base transformer architecture, enabling end-to-end transcription of documents with intrinsic structures, demonstrated on sheet music, shape drawings, and mechanical engineering drawings.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2507.01006","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning","primary_cat":"cs.CV","submitted_at":"2025-07-01T17:55:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"GLM-4.5V reaches state-of-the-art results on 42 multimodal benchmarks among open-source models of similar size by applying reinforcement learning with curriculum sampling to a strong vision foundation model.","context_count":1,"top_context_role":"method","top_context_polarity":"use_method","context_text":"natural images, automatically extracting textual content and their corresponding bounding boxes. The resulting data is subsequently filtered to retain only images containing at least one valid OCR detection, thereby enriching the dataset with authentic, real-world text instances. 3. Academic documents: We adopt a processing methodology inspired by Nougat [5]. A large corpus of papers is sourced from arXiv, where the LaTeX source code is first normalized and converted to HTML format using the LaTeXML tool. The HTML is then parsed and transformed into a lightweight markup language. Finally, this content is segmented according to the original PDF page breaks and rasterized, creating a high-quality dataset of paired PDF page renderings"},{"citing_arxiv_id":"2504.11101","ref_index":3,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Consensus Entropy: Harnessing Multi-VLM Agreement for Self-Verifying and Self-Improving OCR","primary_cat":"cs.CV","submitted_at":"2025-04-15T11:51:18+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Consensus Entropy measures inter-VLM output agreement to verify OCR reliability and enable self-improving ensembles, yielding 42.1% F1 gains over single-model judging.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2410.21169","ref_index":21,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"Document Parsing Unveiled: Techniques, Challenges, and Prospects for Structured Information Extraction","primary_cat":"cs.MM","submitted_at":"2024-10-28T16:11:35+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"Survey proposing a taxonomy for document parsing into pipeline-based systems and VLM-driven unified models, reviewing components, metrics, benchmarks, and challenges.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Extraction [213], [55], [98],ChartDETR [270], [40], [4], [182] FR-DETR [220], [189] End-to-End VLM model General VLMs LLaVA [28, 136], [134, 135] QwenVL [14, 244], InternVL [30, 31], Monkey [142] Specialized VLMs Document Understanding: DocPedia [60], TextMonkey [142], DocOwl [81-83, 277],Vary [253],Fox [133],PDF-Wukong [265] Document Parsing: Nougat [21], Donut [106], GoT [254] Fig. 1. Overview of Document Parsing Methodology. Section 9 discusses current challenges in the field and highlights important future directions. Finally, Section 10 provides a concise and insightful conclusion. The appendix of the survey provides a detailed summary of datasets and metrics related to document parsing. 2 METHODOLOGY"},{"citing_arxiv_id":"2409.18839","ref_index":4,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"MinerU: An Open-Source Solution for Precise Document Content Extraction","primary_cat":"cs.CV","submitted_at":"2024-09-27T15:35:15+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":4.0,"formal_verification":"none","one_line_summary":"MinerU delivers an open-source pipeline for high-precision document content extraction by integrating specialized models with tuned preprocessing and postprocessing rules.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2409.01704","ref_index":6,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model","primary_cat":"cs.CV","submitted_at":"2024-09-03T08:41:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"GOT is a unified end-to-end model that treats all man-made optical signals as characters and handles multiple OCR tasks including formatted output and interactive region recognition via prompts.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2408.16500","ref_index":8,"ref_count":1,"confidence":0.98,"is_internal_anchor":true,"paper_title":"CogVLM2: Visual Language Models for Image and Video Understanding","primary_cat":"cs.CV","submitted_at":"2024-08-29T12:59:12+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"CogVLM2 family achieves state-of-the-art results on image and video understanding benchmarks through improved visual expert architecture, higher resolution inputs, and automated temporal grounding for videos.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2403.05525","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":true,"paper_title":"DeepSeek-VL: Towards Real-World Vision-Language Understanding","primary_cat":"cs.AI","submitted_at":"2024-03-08T18:46:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"DeepSeek-VL develops open-source 1.3B and 7B vision-language models that achieve competitive or state-of-the-art results on real-world visual-language benchmarks through diverse data curation, a hybrid vision encoder, and pretraining that preserves language capabilities.","context_count":1,"top_context_role":"background","top_context_polarity":"unclear","context_text":"Here is how you can translate this flowchart into Python code: In this Python code, we define a function called max_of_twothat takes two arguments, x and y, and returns the maximum of these two values. We then call this function with the appropriate inputs and print the result. PromptPlease generate the latex code of the formula in the image. ResponseA_n= a_0 \\left[ 1 + \\frac{3}{4} \\sum_{k=1}^{n} \\left( \\frac{4}{9} \\right)^k \\right] PromptTell me the main contribution of this paper. ResponseThe main contribution of this paper, as stated in the abstract, is to develop a cognitive model that integrates individual differences into human-automated vehicle interaction. The authors aim to provide insights into how humans interact with new technologies and propose a framework for understanding these interactions."}],"limit":50,"offset":0}