Research map / Architectures and efficiency / RNNs and normalization
CA-RNN: Using Context-Aligned Recurrent Neural Networks for Modeling Sentence Similarity
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 context alignment gating (semantic alignment) to BLSTM with max pooling (ablation) | BLSTM · TREC-QA | MAP: 0.7117 -> 0.8227 | Helped |
| add context alignment gating to BLSTM with attention (ablation) | BLSTM (ATT) · WikiQA | MAP: 0.6979 -> 0.7358 | Helped |
| add context alignment gating to BLSTM with attention (ablation) | BLSTM (ATT) · MSRP | accuracy (%): 74.4 -> 77.3 | Helped |
| use semantic-based instead of overlap-based word alignment (ablation) | CA-RNN · TREC-QA | MAP: 0.8205 -> 0.8227 | Helped |
| use semantic-based instead of overlap-based word alignment (ablation) | CA-RNN (ATT) · MSRP | accuracy (%): 76.5 -> 77.3 | Helped |
| reverse alignment direction to A2Q instead of Q2A (ablation) | CA-RNN-S (MAX) · TREC-QA/WikiQA | MAP | Hurt |
| reverse alignment direction R2L instead of L2R (ablation) | CA-RNN-S (MAX) · MSRP | accuracy (%) | No clear effect |
| replace attention pooling with max pooling in CA-RNN (ablation) | CA-RNN-S · TREC-QA | MAP: 0.8159 -> 0.8227 | No clear effect |
Closest papers by meaning
- Siamese Recurrent Architectures for Learning Sentence Similarity (AAAI 2016 · similarity 0.88)
- Text Alignment Is An Efficient Unified Model for Massive NLP Tasks (NeurIPS 2023 · similarity 0.86)
- LightRNN: Memory and Computation-Efficient Recurrent Neural Networks (NeurIPS 2016 · similarity 0.86)
- Recurrent Hierarchical Topic-Guided RNN for Language Generation (ICML 2020 · similarity 0.86)
- Narrowing the Gap between Supervised and Unsupervised Sentence Representation Learning with Large Language Model (AAAI 2024 · similarity 0.84)
- Weak-to-Strong Search: Align Large Language Models via Searching over Small Language Models (NeurIPS 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.