Research map / Vision-language and multimodal / Visual question answering
Lattice CNNs for Matching Based Chinese Question Answering
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 word-lattice CNNs instead of character-only CNNs | Chinese QA matching · DBQA | MRR: .8809 -> .8902 | Helped |
| Use word-lattice CNNs instead of character-only CNNs | Chinese QA matching · KBRE | MRR: .9570 -> .9592 | Helped |
| Use LCN n-gram context pooling instead of DGC pooling | lattice-based QA matching · DBQA | MRR: .8873 -> .8902 | Helped |
| Combine characters with lattice word information rather than characters alone | CNN matching · DBQA | MRR: .8809 -> .8815 | Helped |
| Replace gated pooling with average pooling (ablation) | LCN · DBQA | MRR: .8902 -> .8869 | Hurt |
| Replace gated pooling with max pooling (ablation) | LCN · DBQA | MRR: .8902 -> .8875 | Hurt |
| Use gated pooling rather than average pooling (ablation) | LCN · KBRE | P@1: 93.60% -> 93.32% | Hurt |
| Add the intersection of CTB and PKU segmented words to a character-based lattice (ablation) | LCN lattice construction · DBQA | MRR: .8809 -> .8851 | Helped |
| Add the union of CTB and PKU segmented words to a character-based lattice (ablation) | LCN lattice construction · DBQA | MRR: .8809 -> .8874 | Helped |
| Add top 10 outputs from each of two segmenters to a character-based lattice (ablation) | LCN lattice construction · DBQA | MRR: .8809 -> .8869 | Helped |
| Use vocabulary-based lattice construction rather than adding segmenter outputs to a character sequence (ablation) | LCN lattice construction · DBQA | MRR: .8869 -> .8902 | Helped |
Closest papers by meaning
- LET: Linguistic Knowledge Enhanced Graph Transformer for Chinese Short Text Matching (AAAI 2021 · similarity 0.88)
- Knowledge and Cross-Pair Pattern Guided Semantic Matching for Question Answering (AAAI 2020 · similarity 0.85)
- Coarse-grain Fine-grain Coattention Network for Multi-evidence Question Answering (ICLR 2019 · similarity 0.85)
- A Compare-Aggregate Model for Matching Text Sequences (ICLR 2017 · similarity 0.85)
- Dynamic Coattention Networks For Question Answering (ICLR 2017 · similarity 0.84)
- Knowledge Enhanced Hybrid Neural Network for Text Matching (AAAI 2018 · 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.