Research map / Architectures and efficiency / RNNs and normalization
Machine Comprehension Using Match-LSTM and Answer Pointer
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 |
|---|---|---|---|
| Predict answer boundaries rather than a sequence of token positions | Match-LSTM with Answer Pointer · SQuAD v1.1 dev | Exact Match: 54.4 -> 61.1 | Helped |
| Predict answer boundaries rather than a sequence of token positions | Match-LSTM with Answer Pointer · SQuAD v1.1 dev | F1: 68.2 -> 71.2 | Helped |
| Add global search over spans no longer than 15 tokens (ablation) | Match-LSTM boundary model · SQuAD v1.1 dev | F1: +1.5% | Helped |
| Increase hidden dimension from 150 to 300 (ablation) | Match-LSTM boundary model with search · SQuAD v1.1 dev | F1: 72.7 -> 72.7 | No clear effect |
| Add bidirectional preprocessing LSTM (ablation) | Match-LSTM boundary model with search · SQuAD v1.1 dev | F1: 72.7 -> 73.0 | No clear effect |
| Add bidirectional Answer Pointer (ablation) | Match-LSTM boundary model with search and bidirectional preprocessing · SQuAD v1.1 dev | F1: 73.0 -> 73.9 | Helped |
| Ensemble five boundary models and search over spans | Match-LSTM boundary model · SQuAD v1.1 test | F1: 73.7 -> 77.0 | Helped |
| Use Match-LSTM with Answer Pointer instead of logistic regression | SQuAD question answering · SQuAD v1.1 dev | Exact Match: 40.0 -> 63.0 | Helped |
Closest papers by meaning
- S-Net: From Answer Extraction to Answer Synthesis for Machine Reading Comprehension (AAAI 2018 · similarity 0.89)
- MacNet: Transferring Knowledge from Machine Comprehension to Sequence-to-Sequence Models (NeurIPS 2018 · similarity 0.87)
- Assessing the Benchmarking Capacity of Machine Reading Comprehension Datasets (AAAI 2020 · similarity 0.87)
- Inference on Syntactic and Semantic Structures for Machine Comprehension (AAAI 2018 · similarity 0.86)
- Retrospective Reader for Machine Reading Comprehension (AAAI 2021 · similarity 0.86)
- Multi-Matching Network for Multiple Choice Reading Comprehension (AAAI 2019 · similarity 0.86)
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.