{"total":14,"items":[{"citing_arxiv_id":"2606.27446","ref_index":60,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Causal Connections: Leveraging Multilingual Fine-Tuning for Financial QA@FinCausal 2026","primary_cat":"cs.CL","submitted_at":"2026-06-25T18:17:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":2.0,"formal_verification":"none","one_line_summary":"Fine-tuned multilingual LLMs achieve top shared-task scores on financial causality extraction in English and Spanish.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.27316","ref_index":51,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"LLM-Based Examination of Eligibility Criteria from Securities Prospectuses at the German Central Bank","primary_cat":"cs.CL","submitted_at":"2026-06-25T17:29:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"LLMs are applied in a generative pipeline for extracting, normalizing, and interpreting eligibility criteria from securities prospectuses, achieving up to 91% precision in document-level decisions with a conservative bias.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.24302","ref_index":11,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"TrOCR for Medieval HTR: A Systematic Ablation Study with Cross-Dataset Validation","primary_cat":"cs.CV","submitted_at":"2026-06-23T08:34:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Systematic ablation of TrOCR fine-tuning for medieval HTR finds that freezing up to three encoder or six decoder layers does not significantly harm accuracy and that removing CLAHE contrast normalization yields comparable 7.84% CER on the Cortonese manuscript.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2606.02162","ref_index":10,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Multimodal Approaches for Visually-Rich Document Type Classification: A Comparative Analysis","primary_cat":"cs.CV","submitted_at":"2026-06-01T12:24:26+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":2.0,"formal_verification":"none","one_line_summary":"Specialized multimodal transformers outperform LLM-based models on visually rich documents, with image information contributing most to classification accuracy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.19866","ref_index":2,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"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":"2605.17447","ref_index":11,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"FastOCR: Dynamic Visual Fixation via KV Cache Pruning for Efficient Document Parsing","primary_cat":"cs.CV","submitted_at":"2026-05-17T13:39:47+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"FastOCR dynamically selects a small subset of visual tokens per decoding step using focal-guided pruning and cross-step reuse, retaining 98% accuracy on Qwen2.5-VL while attending to only 5% of tokens and cutting attention latency by 3x.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.17159","ref_index":17,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"MADP: A Multi-Agent Pipeline for Sustainable Document Processing with Human-in-the-Loop","primary_cat":"cs.AI","submitted_at":"2026-05-16T21:18:39+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"MADP multi-agent pipeline with human-in-the-loop achieves 97% full automation on 955 real documents, 98.5% accuracy on ablation set, and 69-70% reductions in FTE, energy, and emissions versus manual processing.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16504","ref_index":5,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"From Handwriting to Structured Data: Benchmarking AI Digitisation of Handwritten Forms","primary_cat":"cs.CV","submitted_at":"2026-04-14T17:45:22+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"Frontier multimodal LLMs achieve ~85% accuracy and ~90% weighted F1 on digitizing complex handwritten medical forms, with Gemini 3.1 strongest overall and prompt optimization lifting macro metrics over 60%.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.04771","ref_index":16,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"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":"This modular design enables independent optimization of each component but suffers from error propagation and inter-module information loss. 4 MinerU2.5-Pro: Pushing the Limits of Data-Centric Document Parsing at Scale End-to-end VLM methods.These methods directly map document images to structured output, avoiding the cascading errors inherent in pipeline approaches. Nougat [2], built on the Donut architec- ture [16], established a strong baseline for the image-to-markup paradigm on academic documents; GOT-OCR 2.0 [39] unified scene text and document OCR within a single model. Subsequent works such as Ocean-OCR [3], olmOCR [29], and dots.ocr [18] employ native-resolution visual encoders to further improve performance. However, native-resolution processing incurs O(N 2) token complexity,"},{"citing_arxiv_id":"2604.03476","ref_index":3,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"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":"2603.19790","ref_index":21,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"From Plausibility to Verifiability: Risk-Controlled Generative OCR with Vision-Language Models","primary_cat":"cs.CV","submitted_at":"2026-03-20T09:28:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A model-agnostic Geometric Risk Controller reduces extreme errors in VLM-based OCR by requiring cross-view consensus before accepting outputs.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2602.01785","ref_index":52,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"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":"background","top_context_polarity":"background","context_text":"15664 [cs.LG] https://arxiv.org/abs/2111.15664 [51] Kenton Lee, Mandar Joshi, Iulia Turc, Hexiang Hu, Fangyu Liu, Julian Eisenschlos, Urvashi Khandelwal, Peter Shaw, Ming-Wei Chang, and Kristina Toutanova. 2023. Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding. arXiv:2210.03347 [cs.CL] https://arxiv.org/abs/2210.03347 [52] Han Li, Yuling Shi, Shaoxin Lin, Xiaodong Gu, Heng Lian, Xin Wang, Yantao Jia, Tao Huang, and Qianxiang Wang. 2025. Swe-debate: Competitive multi-agent debate for software issue resolution. arXiv:2507.23348 [cs.SE] [53] Minghao Li, Tengchao Lv, Jingye Chen, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, and Furu Wei. 2022. TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models."},{"citing_arxiv_id":"2502.20295","ref_index":21,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Judge a Book by its Cover: Investigating Multi-Modal LLMs for Multi-Page Handwritten Document Transcription","primary_cat":"cs.LG","submitted_at":"2025-02-27T17:21:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Introduces OCR+PAGE-1 and OCR+PAGE-N prompting strategies that improve zero-shot multi-page handwritten document transcription by sharing context across pages.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2308.13418","ref_index":30,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Nougat: Neural Optical Understanding for Academic Documents","primary_cat":"cs.LG","submitted_at":"2023-08-25T15:03:36+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Nougat applies a visual transformer to convert academic PDFs into markup language while accurately handling mathematical content on a new scientific document dataset.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}