Research map / Architectures and efficiency / Attention and transformers
SG-Net: Syntax-Guided Machine Reading Comprehension
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 syntax-guided self-attention with dual-context aggregation | BERT whole word masking · SQuAD 2.0 dev | EM: 84.1 -> 85.1 | Helped |
| Add syntax-guided self-attention with dual-context aggregation | BERT whole word masking · SQuAD 2.0 dev | F1: 86.8 -> 87.9 | Helped |
| Add syntax-guided self-attention with dual-context aggregation | BERT whole word masking · RACE | accuracy: 72.6 -> 74.2 | Helped |
| Add an answerability verifier | SG-Net · SQuAD 2.0 test | EM: 85.1 -> 85.2 | Helped |
| Add an answerability verifier | SG-Net · SQuAD 2.0 test | F1: 87.9 -> 87.9 | Helped |
| Use an added vanilla attention layer without syntax guidance or dual-context aggregation (ablation) | BERT whole word masking · SQuAD 2.0 dev | EM: 84.1 -> 84.2 | No clear effect |
| Use syntax-guided attention only, without vanilla attention in dual-context aggregation (ablation) | BERT whole word masking · SQuAD 2.0 dev | EM: 84.1 -> 84.4 | Helped |
| Add vanilla attention representations to syntax-guided attention through dual-context aggregation (ablation) | Syntax-guided attention only · SQuAD 2.0 dev | EM: 84.4 -> 85.1 | Helped |
| Replace weighted dual aggregation with concatenation (ablation) | SG-Net · SQuAD 2.0 dev | EM: 85.1 -> 84.5 | Hurt |
| Replace weighted dual aggregation with bi-attention (ablation) | SG-Net · SQuAD 2.0 dev | EM: 85.1 -> 84.9 | Hurt |
Closest papers by meaning
- MacNet: Transferring Knowledge from Machine Comprehension to Sequence-to-Sequence Models (NeurIPS 2018 · similarity 0.87)
- A Graph Fusion Approach for Cross-Lingual Machine Reading Comprehension (AAAI 2023 · similarity 0.86)
- Zero-Shot Cross-Lingual Machine Reading Comprehension via Inter-sentence Dependency Graph (AAAI 2022 · similarity 0.86)
- SAS: Simple Attention Sparsification via End-to-End Optimization of Context Ranking (NeurIPS 2026 · similarity 0.84)
- TAIA: Large Language Models are Out-of-Distribution Data Learners (NeurIPS 2024 · similarity 0.84)
- Mastering Symbolic Operations: Augmenting Language Models with Compiled Neural Networks (ICLR 2024 · 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.