Research map / Vision-language and multimodal / Visual question answering
EA Reader: Enhance Attentive Reader for Cloze-Style Question Answering via Multi-Space Context Fusion
Chengzhen Fu, Yan Zhang
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 |
|---|---|---|---|
| add multi-space projections (ablation) | MSCF · CNN | test accuracy: 73.4 -> 79.3 | Improved |
| add multi-space projections (ablation) | MSCF · CBT-CN | test accuracy: 68.5 -> 71.5 | Improved |
| use bit-wise attention instead of token-wise bilinear (ablation) | MSCF · CNN | test accuracy: 76.5 -> 79.3 | Improved |
| use bit-wise attention instead of token-wise bilinear (ablation) | MSCF · Daily Mail | test accuracy: 80.4 -> 82.7 | Improved |
| use bit-wise attention instead of token-wise bilinear (ablation) | MSCF · CBT-CN | test accuracy: 69.5 -> 71.5 | Improved |
| multi-space projections alone vs Stanford AR | MSCF w/o bit-wise · CNN | test accuracy: 72.4 -> 76.5 | Improved |
| add gated write operation to multi-hop reader (ablation) | EA Reader · Daily Mail | test accuracy: 83.4 -> 84.3 | Improved |
| add gated write operation to multi-hop reader (ablation) | EA Reader · CBT-NE | test accuracy: 75.1 -> 76.6 | Improved |
| extend single-hop MSCF to multi-hop EA Reader | EA Reader · CNN | test accuracy: 79.3 -> 81.4 | Improved |
| extend single-hop MSCF to multi-hop EA Reader | EA Reader · Daily Mail | test accuracy: 82.7 -> 84.3 | Improved |
| bit-wise vs multi-head attention scoring | MSCF single-hop · CNN | test accuracy: 77.8 -> 79.3 | Improved |
| vary number of latent subspaces beyond 6 | EA Reader · CNN, CBT-NE | validation accuracy | Hurt |
Closest papers by meaning
- Autofocus Retrieval: An Effective Pipeline for Multi-Hop Question Answering With Semi-Structured Knowledge (NeurIPS 2026 · similarity 0.87)
- Cooperative Retrieval-Augmented Generation for Question Answering: Mutual Information Exchange and Ranking by Contrasting Layers (NeurIPS 2025 · similarity 0.87)
- DSAS: A Universal Plug-and-Play Framework for Attention Optimization in Multi-Document Question Answering (NeurIPS 2025 · similarity 0.86)
- The Atlas of In-Context Learning: How Attention Heads Shape In-Context Retrieval Augmentation (NeurIPS 2025 · similarity 0.85)
- Making Retrieval-Augmented Language Models Robust to Irrelevant Context (ICLR 2024 · similarity 0.85)
- Bridging Context Gaps: Leveraging Coreference Resolution for Long Contextual Understanding (ICLR 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.
