Summary of Doge: Towards Versatile Visual Document Grounding and Referring, by Yinan Zhou et al.
DOGE: Towards Versatile Visual Document Grounding and Referring
by Yinan Zhou, Yuxin Chen, Haokun Lin, Shuyu Yang, Li Zhu, Zhongang Qi, Chen Ma, Ying Shan
First submitted to arxiv on: 26 Nov 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed DOcument Grounding and Eferring data engine (DOGE-Engine) aims to bridge the gap in multimodal large language models (MLLMs) for visual document understanding by producing high-quality fine-grained document data. The engine generates multi-granular parsing data for text localization and recognition, as well as instruction-tuning data to enhance grounding and referring capabilities during dialogue and reasoning. A comprehensive benchmark, DOGE-Bench, is constructed with 7 tasks across 3 document types (chart, poster, PDF document), providing evaluations for fine-grained document understanding. The baseline model, DOGE, is developed using the generated data, demonstrating accurate text referral and grounding at multiple granularities within document images. This pioneering MLLM will be open-sourced for community development. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new tool to help computers better understand documents with pictures or charts. It creates special data that helps machines learn to find specific words in these documents and understand what they mean. The tool also tests how well machines can do this job, using different types of documents like charts and posters. The goal is to make machines smarter at understanding documents with visual information. |
Keywords
» Artificial intelligence » Grounding » Instruction tuning » Parsing
