← Back to all news
arXiv AI By Zhuoran Yu, Le Thien Phuc Nguyen, Jaden Park, Xinyi Gu, Zexue He, Soochahn Lee, Rogerio Feris, Yong Jae Lee

DocHop: Benchmarking Out-of-domain Multi-hop Reasoning in Information-Dense Documents

Read the original on arXiv AI →

DocHop is a new benchmark that tests multimodal large language models on integrated chart‑context reasoning within document‑style images. The benchmark presents narrative text that imposes multi‑step compositional constraints, while charts supply the data needed to answer questions grounded in semantic reference labels. It contains 2,074 examples across six task categories, generated via a stochastic logic‑first pipeline that controls reasoning depth and visual density, and shows a large performance gap between humans (over 90% accuracy) and the best models (62.83%).

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 AI.

Related stories

arXiv AI
3d ago

DocPilot: Interactive Agent Orchestration and Training for Long Visual Document Understanding

DocPilot is an agentic framework for long visual document understanding that orchestrates evidence seeking through page-level navigation and fine-grained localization, integrating visual inspection, code-based numerical reasoning, and cross-modal reasoning. It employs curriculum-based agentic training to enable small language models to learn complex document solving, progressing from single-ability learning to adaptive orchestration. Experiments on benchmarks such as MMLongBench-Doc, DocBench, FinRAGBench-V, and MADQA show that DocPilot achieves strong gains with frontier VLM backbones, and its 4B variant competes with much larger models, demonstrating effectiveness and generalization across diverse long-document reasoning tasks.

By Zhuowen Liang, Xiaotian Lin, Yuyu Luo, Haixun Wang, Nan Tang
llmsagentsnlpmultimodal
More like this →
arXiv AI
Aug 5

DocTrace: Towards Traceable Long Document VQA via Hierarchical Evidence Graph Reasoning

arXiv:2608. 03292v1 Announce Type: new Abstract: Long Document Visual Question Answering (LongDocVQA) requires Multimodal Large Language Models (MLLMs) to locate, integrate, and reason over heterogeneous document elements distributed across multiple pages.

By Le Xiang, Zhicheng Guan, Hong Chen, Xiaocong Lin, Zhenghua Lei, Teng Hu, Bolei He, Long Zeng
llmsragagentsnlpfine-tuningmultimodalbenchmarks
More like this →
arXiv Computation and Language
Sep 1

CoVA-SFT: A Large-Scale Dataset for Chain of Visual Abstractions

CoVA‑SFT is a new large‑scale dataset comprising 51.9K samples and over 222K multimodal reasoning steps that teach models to interleave text and visual abstractions across five layout families and 17 complex tasks. It includes explicit rationale formulations, agentic renderings, and verification loops to help models build and maintain internal visual workspaces for purely textual reasoning problems. A companion benchmark, CoVA‑Bench, contains 1,700 held‑out test samples for reproducible evaluation, and models fine‑tuned on CoVA‑SFT outperform all interleaved CoT baselines by more than 2× on average, though they still lag behind strong text‑only CoT baselines.

By Tsung-Han Wu, Heekyung Lee, Anya Ji, Haoming Chen, Trevor Darrell, Joseph E. Gonzalez, David M. Chan
llmsagentsfine-tuningmultimodalbenchmarks
More like this →