Research map / Architectures and efficiency / CNN design
Understanding the Semantic Structures of Tables with a Hybrid Deep Neural Network Architecture
Kyosuke Nishida, Kugatsu Sadamitsu, Ryuichiro Higashinaka, Yoshihiro Matsuo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| replace handcrafted-feature Random Forest (Eberius15) with TabNet (RNN+CNN) | table type classification · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1: 0.8165 -> 0.8842 ± 0.0070 | Improved |
| replace Bidirectional HAN with TabNet's CNN over cell matrix | deep table encoder · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1: 0.8562 ± 0.0045 -> 0.8842 ± 0.0070 | Improved |
| ensemble five TabNets by majority voting instead of a single model | TabNet · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1: 0.8842 -> 0.9105 | Improved |
| ensemble five HANs instead of a single HAN | HAN · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1: 0.8409 -> 0.8471 | Improved |
| use more training websites (1 to 300; 29,050 to 60,678 tables) | TabNet and Bidirectional HAN · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1 | Improved |
| add pooling operations in the CNN (ablation) | TabNet · Web tables test (3,567 tables, 200 unseen websites) | weighted macro-averaged F1 | Hurt |
Closest papers by meaning
- Interpretable Mesomorphic Networks for Tabular Data (NeurIPS 2024 · similarity 0.87)
- Making Pre-trained Language Models Great on Tabular Prediction (ICLR 2024 · similarity 0.86)
- TabPedia: Towards Comprehensive Visual Table Understanding with Concept Synergy (NeurIPS 2024 · similarity 0.85)
- Table as a Modality for Large Language Models (NeurIPS 2025 · similarity 0.85)
- TabGLM: Tabular Graph Language Model for Learning Transferable Representations Through Multi-Modal Consistency Minimization (AAAI 2025 · similarity 0.85)
- Hybrid Autoencoders for Tabular Data: Leveraging Model-Based Augmentation in Low-Label Settings (NeurIPS 2025 · similarity 0.84)
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.
