Research map / Vision-language and multimodal / Human and embodied modeling
SelfDoc: Self-Supervised Document Representation Learning
Check 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 modality-adaptive attention (ablation) | SelfDoc · FUNSD test | micro F1: 0.8263 -> 0.8336 | Helped |
| Add modality-adaptive attention (ablation) | SelfDoc · RVL-CDIP test | accuracy: 0.9263 -> 0.9281 | Helped |
| Pre-train on documents rather than train from scratch (ablation) | SelfDoc without M-AA · FUNSD test | micro F1: 0.7607 -> 0.8263 | Helped |
| Pre-train on documents rather than train from scratch (ablation) | SelfDoc without M-AA · RVL-CDIP test | accuracy: 0.9049 -> 0.9263 | Helped |
| Add layout input (ablation) | SelfDoc trained from scratch · RVL-CDIP test | accuracy: 0.9070 -> 0.9049 | Helped |
| Add layout input (ablation) | SelfDoc trained from scratch · FUNSD test | micro F1: 0.7579 -> 0.7607 | Helped |
| Add language input (ablation) | SelfDoc trained from scratch · FUNSD test | micro F1: 0.6209 -> 0.7607 | Helped |
| Add language input (ablation) | SelfDoc trained from scratch · RVL-CDIP test | accuracy: 0.8447 -> 0.9049 | Helped |
| Add vision input (ablation) | SelfDoc trained from scratch · FUNSD test | micro F1: 0.7491 -> 0.7607 | Helped |
| Add vision input (ablation) | SelfDoc trained from scratch · RVL-CDIP test | accuracy: 0.8895 -> 0.9049 | Helped |
| Increase pre-training data from 40K to 320K (ablation) | SelfDoc · FUNSD test | micro F1: 0.7886 -> 0.8263 | Helped |
| Increase pre-training data from 40K to 320K (ablation) | SelfDoc · RVL-CDIP test | accuracy: 0.9119 -> 0.9263 | Helped |
Closest papers by meaning
- DocMamba: Efficient Document Pre-training with State Space Model (AAAI 2025 · similarity 0.85)
- M2Doc: A Multi-Modal Fusion Approach for Document Layout Analysis (AAAI 2024 · similarity 0.85)
- InstructDoc: A Dataset for Zero-Shot Generalization of Visual Document Understanding with Instructions (AAAI 2024 · similarity 0.85)
- Enhancing Visual Document Understanding with Contrastive Learning in Large Visual-Language Models (CVPR 2024 · similarity 0.85)
- ITO: Multi-View Alignment and Training-Time Fusion for Image-Text Pretraining (NeurIPS 2026 · similarity 0.85)
- Recurrence-Enhanced Vision-and-Language Transformers for Robust Multimodal Document Retrieval (CVPR 2025 · similarity 0.85)
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.