arXiv Machine Learning October 16, 2026 By Yong Zu, Ninghua Yang, Zezhong Tan, Xinhong Ma, Ziqiang Dong, Yang Yang

P2TPO: Structure-Aligned Page-to-Token Credit Assignment for Full-Page Document Parsing

Read the original on arXiv Machine Learning →

P2TPO introduces a structure‑aligned credit‑assignment framework for full‑page document parsing, aligning block‑level quality contrasts with generated spans and redistributing credit within matched blocks. The method employs Unified Anchor Pricing to handle matched and missing elements and Span Credit Redistribution to adjust token‑level credit while preserving overall block advantage. Experiments show P2TPO surpasses existing policy optimization baselines across six parsing benchmarks and scales well with model size and initialization.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

Sep 10

AtomCite: Verification and Correction of Supplied Page-Level Citations in Multi-Page Documents

AtomCite is an agentic framework that verifies and corrects page‑level citations in multi‑page documents by parsing answers into claims, checking each claim against the cited page image, and applying a deterministic repair policy. The authors introduce DocCite, the first benchmark for this task, built on MP‑DocVQA and DUDE, containing 928 injected instances and 1,909 verified natural errors. Across Gemini, Claude, and GPT models, AtomCite achieves about 93% verification accuracy and improves citation precision from 34% to 87‑90%, while also enhancing hallucination detection in open‑source models.

By Chen Qian, Yimeng Wang, Yu Chen, Lingfei Wu, Andreas Stathopoulos

llms agents computer-vision benchmarks safety

arXiv Machine Learning

Oct 22

One Act, Many Masks: Effect-Marginal Policy Optimization for Long-Horizon Agents

The paper introduces Effect-Marginal Policy Optimization (EMPO), a method that addresses credit-update misalignment in multi-turn reinforcement learning for large language model agents. EMPO aligns policy updates with execution-level equivalence by using an effect-level surrogate that couples updates across different textual responses that produce the same effect. Experiments on ALFWorld, WebShop, and Sokoban demonstrate that EMPO improves both task performance and learning efficiency compared to strong baselines.

By Shaojin Ma, Ruitao Wang, Pengyi Li, Jie Yang, Liangyu Chen, Gangyi Zhao, Hongyao Tang, Yan Zheng, Jianye Hao

llms agents reinforcement-learning

Hugging Face Trending Papers

Jul 21

HPD-Parsing: Hierarchical Parallel Document Parsing

Efficient teamwork typically combines global coordination with parallel execution, a principle not yet fully reflected in unified Vision-Language Model (VLM)-based document parsers. Existing unified parsers process an entire page jointly but generate its output through a single token-by-token autoregressive trajectory, creating a sequential bottleneck that grows with document length.

llms efficiency multimodal benchmarks

arXiv AI

Jul 16

OvisOCR2 Technical Report

arXiv:2607. 13639v1 Announce Type: cross Abstract: We introduce OvisOCR2, a 0.

By Shiyin Lu, Yinglun Li, Yu Xia, Yuhui Chen, An-Yang Ji, Jun-Peng Jiang, Qing-Guo Chen, Jianshan Zhao, En Lin, Haijun Li, Cheng Qin, Zhao Xu, Weihua Luo

reinforcement-learning fine-tuning efficiency benchmarks

arXiv Computation and Language

Sep 17

How Do Document Parsers Break? Auditing Structural Vulnerability in Document Intelligence

arXiv:2605.19309v4 Announce Type: replace Abstract: Document Layout Analysis (DLA) pipelines provide structured page representations for retrieval-augmented generation, long-document question answeri...

By Yue Chen, Yihao Wang, Ziyi Tang, Yongsen Zheng, Keze Wang

rag computer-vision nlp safety

arXiv Computer Vision

Sep 18

WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing

WeVisDoc is a two‑stage data‑centric framework designed to improve end‑to‑end document parsing. Stage I expands coverage by adding heterogeneous data and applying structure‑preserving degradation synthesis, while Stage II evaluates residual errors with a held‑out probe and uses those diagnostics to target data construction and token budget reallocation. The resulting WeVisDoc‑4B model achieves an overall score of 95.38 on OmniDocBench v1.6 and outperforms competing parsers across all evaluated settings, with Stage II delivering notable gains on degraded tracks.

By Hao Yu, Kang Liu, Linnan Zhao, Jiabo Zhan, Chong Sun, Chen Li, Jing Lyu

benchmarks