Research map / Vision-language and multimodal / Human and embodied modeling
Words or Characters? Fine-grained Gating for 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 |
|---|---|---|---|
| use feature-conditioned fine-grained vector gating instead of scalar gating (ablation) | word-character representations on Twitter · Twitter test | Precision@1: 0.2974 -> 0.3069 | Helped |
| use feature-conditioned fine-grained vector gating instead of scalar gating (ablation) | word-character representations on Twitter · Twitter test | Recall@10: 0.4982 -> 0.5119 | Helped |
| use feature-conditioned fine-grained vector gating instead of scalar gating (ablation) | word-character representations on Twitter · Twitter test | Mean Rank: 104.2 -> 101.5 | Helped |
| use token features to compute the gate rather than concatenate them to the representations (ablation) | word-character representations on Twitter · Twitter test | Precision@1: 0.2951 -> 0.3069 | Helped |
| use fine-grained document-query gating instead of gated attention (ablation) | CBT reading comprehension model · CBT CN test | accuracy: 0.7084 -> 0.7204 | Helped |
| use fine-grained word-character gating instead of concatenation (ablation) | CBT reading comprehension model · CBT CN test | accuracy: 0.696 -> 0.7204 | Helped |
| use fine-grained document-query gating instead of gated attention (ablation) | SQuAD reading comprehension model · SQuAD dev | F1: 0.6983 -> 0.7125 | Helped |
| use fine-grained word-character gating instead of concatenation (ablation) | SQuAD reading comprehension model · SQuAD dev | F1: 0.6857 -> 0.7125 | Helped |
| use fine-grained document-query gating instead of gated attention (ablation) | CBT reading comprehension model · CBT NE dev | accuracy: 0.7890 -> 0.7910 | Helped |
Closest papers by meaning
- Neural Symbolic Reader: Scalable Integration of Distributed and Symbolic Representations for Reading Comprehension (ICLR 2020 · similarity 0.86)
- Generating Distractors for Reading Comprehension Questions from Real Examinations (AAAI 2019 · similarity 0.86)
- Co-Attention Hierarchical Network: Generating Coherent Long Distractors for Reading Comprehension (AAAI 2020 · similarity 0.84)
- Hierarchical Attention Flow for Multiple-Choice Reading Comprehension (AAAI 2018 · similarity 0.84)
- Using Next Sentence Prediction to Test ChatGPT’s Text Comprehension (Student Abstract) (AAAI 2025 · similarity 0.83)
- Thought-Path Contrastive Learning via Premise-Oriented Data Augmentation for Logical Reading Comprehension (AAAI 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.