Research map / Vision-language and multimodal / Visual question answering
Read + Verify: Machine Reading Comprehension with Unanswerable Questions
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 the independent span loss (ablation) | RMR reader · SQuAD 2.0 dev, HasAns | F1: 80.4 -> 81.6 | Helped |
| Add the independent no-answer loss (ablation) | RMR reader · SQuAD 2.0 dev, NoAns | ACC: 69.8 -> 73.1 | Helped |
| Add both independent auxiliary losses (ablation) | RMR reader · SQuAD 2.0 dev, All | F1: 67.5 -> 69.1 | Helped |
| Add the independent span loss (ablation) | RMR + ELMo reader · SQuAD 2.0 dev, HasAns | F1: 86.5 -> 86.8 | Helped |
| Add the independent no-answer loss (ablation) | RMR + ELMo reader · SQuAD 2.0 dev, NoAns | ACC: 75.1 -> 77.0 | Helped |
| Add both independent auxiliary losses (ablation) | RMR + ELMo reader · SQuAD 2.0 dev, All | F1: 71.9 -> 73.7 | Helped |
| Combine the sequential and interactive verifier architectures (ablation) | Answer verifier · SQuAD 2.0 dev, NoAns | ACC: 74.5 / 74.6 -> 76.2 | Helped |
| Add ELMo embeddings to Verifier Model 3 (ablation) | RMR + Verifier Model 3 · SQuAD 2.0 dev, All | F1: 71.5 -> 71.2 | Hurt |
| Add Verifier Model 3 to RMR | RMR · SQuAD 2.0 dev, NoAns | ACC: 73.1 -> 77.1 | Helped |
| Add Verifier Model 3 to RMR + ELMo | RMR + ELMo · SQuAD 2.0 dev, All | F1: 73.7 -> 74.8 | Helped |
| Add Verifier Model 3 to DocQA | DocQA · SQuAD 2.0 dev, NoAns | ACC: 69.1 to 75.2 | Helped |
| Add Verifier Model 3 to DocQA + ELMo | DocQA + ELMo · SQuAD 2.0 dev, NoAns | ACC: 70.6 to 76.1 | Helped |
Closest papers by meaning
- Retrospective Reader for Machine Reading Comprehension (AAAI 2021 · similarity 0.89)
- S-Net: From Answer Extraction to Answer Synthesis for Machine Reading Comprehension (AAAI 2018 · similarity 0.87)
- NeurQuRI: Neural Question Requirement Inspector for Answerability Prediction in Machine Reading Comprehension (ICLR 2020 · similarity 0.86)
- GAAMA 2.0: An Integrated System That Answers Boolean and Extractive Questions (AAAI 2023 · similarity 0.86)
- Generating Well-Formed Answers by Machine Reading with Stochastic Selector Networks (AAAI 2020 · similarity 0.85)
- SuRe: Summarizing Retrievals using Answer Candidates for Open-domain QA of LLMs (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.