Research map / Large language models / Retrieval and knowledge
DCMN+: Dual Co-Matching Network for Multi-Choice 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 bidirectional matching among passage, question, and answer option (ablation) | BERT_base · RACE development | accuracy: 64.6 -> 67.1 | Helped |
| Add passage sentence selection to DCMN (ablation) | BERT_base · RACE development | accuracy: 66.0 -> 66.6 | Helped |
| Add answer option interaction to DCMN (ablation) | BERT_base · RACE development | accuracy: 66.0 -> 66.8 | Helped |
| Add passage sentence selection and answer option interaction to DCMN (ablation) | BERT_base · RACE development | accuracy: 66.0 -> 67.4 | Helped |
| Add passage sentence selection to DCMN (ablation) | BERT_large · RACE development | accuracy: 73.8 -> 74.6 | Helped |
| Add answer option interaction to DCMN (ablation) | BERT_large · RACE development | accuracy: 73.8 -> 74.4 | Helped |
| Add both reading-strategy components to DCMN (ablation) | BERT_large · RACE development | accuracy: 73.8 -> 75.4 | Helped |
| Add passage sentence selection to DCMN (ablation) | XLNet_large · RACE development | accuracy: 81.5 -> 82.1 | Helped |
| Add answer option interaction to DCMN (ablation) | XLNet_large · RACE development | accuracy: 81.5 -> 82.2 | Helped |
| Add both reading-strategy components to DCMN (ablation) | XLNet_large · RACE development | accuracy: 81.5 -> 82.6 | Helped |
| Use cosine rather than bilinear sentence scoring (ablation) | Passage sentence selection with BERT_base · COIN development | accuracy: 82.8 -> 83.5 | Helped |
| Use bilinear rather than cosine sentence scoring (ablation) | Passage sentence selection with BERT_base · RACE development | accuracy: 66.5 -> 66.8 | Helped |
Closest papers by meaning
- Multi-Matching Network for Multiple Choice Reading Comprehension (AAAI 2019 · similarity 0.92)
- Hierarchical Attention Flow for Multiple-Choice Reading Comprehension (AAAI 2018 · similarity 0.89)
- Reverse Multi-Choice Dialogue Commonsense Inference with Graph-of-Thought (AAAI 2024 · similarity 0.86)
- Co-Attention Hierarchical Network: Generating Coherent Long Distractors for Reading Comprehension (AAAI 2020 · similarity 0.86)
- Answer, Assemble, Ace: Understanding How LMs Answer Multiple Choice Questions (ICLR 2025 · similarity 0.84)
- Listening to the Wise Few: Query–Key Alignment Unlocks Latent Correct Answers in Large Language Models (NeurIPS 2026 · 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.