Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.01832 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911189003403264 |
|---|---|
| author | Liu, Shicheng Sun, Kai Fu, Lisheng Chen, Xilun Zhang, Xinyuan Lin, Zhaojiang Shao, Rulin Liu, Yue Kumar, Anuj Yih, Wen-tau Dong, Xin Luna |
| author_facet | Liu, Shicheng Sun, Kai Fu, Lisheng Chen, Xilun Zhang, Xinyuan Lin, Zhaojiang Shao, Rulin Liu, Yue Kumar, Anuj Yih, Wen-tau Dong, Xin Luna |
| contents | Semi-structured content in HTML tables, lists, and infoboxes accounts for a substantial share of factual data on the web, yet the formatting complicates usage, and reliably extracting structured information from them remains challenging. Existing methods either lack generalization or are resource-intensive due to per-page LLM inference. In this paper, we introduce SCRIBES (SCRIpt-Based Semi-Structured Content Extraction at Web-Scale), a novel reinforcement learning framework that leverages layout similarity across webpages within the same site as a reward signal. Instead of processing each page individually, SCRIBES generates reusable extraction scripts that can be applied to groups of structurally similar webpages. Our approach further improves by iteratively training on synthetic annotations from in-the-wild CommonCrawl data. Experiments show that our approach outperforms strong baselines by over 13% in script quality and boosts downstream question answering accuracy by more than 4% for GPT-4o, enabling scalable and resource-efficient web information extraction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_01832 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | SCRIBES: Web-Scale Script-Based Semi-Structured Data Extraction with Reinforcement Learning Liu, Shicheng Sun, Kai Fu, Lisheng Chen, Xilun Zhang, Xinyuan Lin, Zhaojiang Shao, Rulin Liu, Yue Kumar, Anuj Yih, Wen-tau Dong, Xin Luna Computation and Language Semi-structured content in HTML tables, lists, and infoboxes accounts for a substantial share of factual data on the web, yet the formatting complicates usage, and reliably extracting structured information from them remains challenging. Existing methods either lack generalization or are resource-intensive due to per-page LLM inference. In this paper, we introduce SCRIBES (SCRIpt-Based Semi-Structured Content Extraction at Web-Scale), a novel reinforcement learning framework that leverages layout similarity across webpages within the same site as a reward signal. Instead of processing each page individually, SCRIBES generates reusable extraction scripts that can be applied to groups of structurally similar webpages. Our approach further improves by iteratively training on synthetic annotations from in-the-wild CommonCrawl data. Experiments show that our approach outperforms strong baselines by over 13% in script quality and boosts downstream question answering accuracy by more than 4% for GPT-4o, enabling scalable and resource-efficient web information extraction. |
| title | SCRIBES: Web-Scale Script-Based Semi-Structured Data Extraction with Reinforcement Learning |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2510.01832 |