Research map / Vision-language and multimodal / Human and embodied modeling
GeoLayoutLM: Geometric Pre-Training for Visual Information Extraction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add geometric pre-training to document pre-training (ablation) | GeoLayoutLM · FUNSD RE | F1: 74.91 -> 85.32 | Helped |
| add GeoPair geometric pre-training (ablation) | GeoLayoutLM · FUNSD SER | SER: 83.39 -> 91.80 | Helped |
| add GeoMPair geometric pre-training (ablation) | GeoLayoutLM · FUNSD RE | RE: 74.91 -> 82.56 | Helped |
| add GeoTriplet geometric pre-training (ablation) | GeoLayoutLM · FUNSD RE | RE: 74.91 -> 78.90 | Helped |
| combine GeoPair and GeoMPair pre-training (ablation) | GeoLayoutLM · FUNSD RE | RE: 82.23 -> 85.22 | Helped |
| combine all three geometric pre-training tasks (ablation) | GeoLayoutLM · FUNSD | SER and RE: 91.86 -> 92.17; 85.22 -> 85.32 | Helped |
| pre-train the CRP head (ablation) | GeoLayoutLM · FUNSD RE | F1: 82.2 -> 84.9 | Helped |
| add an unpretrained RFE head to pre-trained CRP (ablation) | GeoLayoutLM · FUNSD RE | F1: 84.9 -> 84.0 | Hurt |
| pre-train the RFE head (ablation) | GeoLayoutLM · FUNSD RE | F1: 84.0 -> 86.2 | Helped |
| pre-train both CRP and RFE heads (ablation) | GeoLayoutLM · FUNSD RE | F1: 84.0 -> 86.9 | Helped |
| add RSF postprocessing (ablation) | GeoLayoutLM · FUNSD RE | F1: 87.64 -> 88.62 | Helped |
| add RSF variance loss without postprocessing (ablation) | GeoLayoutLM · FUNSD RE | F1: 87.64 -> 87.62 | No clear effect |
Closest papers by meaning
- Relation-Rich Visual Document Generator for Visual Information Extraction (CVPR 2025 · similarity 0.87)
- LDP: Generalizing to Multilingual Visual Information Extraction by Language Decoupled Pretraining (AAAI 2025 · similarity 0.85)
- Reverse Region-to-Entity Annotation for Pixel-Level Visual Entity Linking (AAAI 2025 · similarity 0.84)
- Prototype-Guided Multimodal Relation Extraction based on Entity Attributes (AAAI 2025 · similarity 0.84)
- Unaligned Message-Passing and Contextualized-Pretraining for Robust Geo-Entity Resolution (AAAI 2025 · similarity 0.83)
- GeoX: Geometric Problem Solving Through Unified Formalized Vision-Language Pre-training (ICLR 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.