Research map / Architectures and efficiency / RNNs and normalization
Cross Temporal Recurrent Networks for Ranking Question Answer Pairs
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 cross-applied question-answer temporal gates to QRNN | QRNN · YahooQA | P@1: 0.573 -> 0.601 | Helped |
| Add cross-applied question-answer temporal gates to QRNN | QRNN · YahooQA | MRR: 0.736 -> 0.755 | Helped |
| Add cross-applied question-answer temporal gates to QRNN | QRNN · QatarLiving | P@1: QRNN by $3\%$ on P@1 | Helped |
| Add cross-applied question-answer temporal gates to QRNN | QRNN · QatarLiving | MAP | Helped |
| Add cross-applied question-answer temporal gates to QRNN | QRNN · TrecQA TRAIN | MAP | Helped |
| Add cross-applied question-answer temporal gates to QRNN | QRNN · TrecQA TRAIN-ALL | MAP | Helped |
| Use CTRN rather than an attention-based matcher | YahooQA · YahooQA | P@1: about $2\%-3\%$ | Helped |
| Use CTRN rather than AI-CNN | QatarLiving · QatarLiving | P@1: $2.5\%$ | Helped |
| Use CTRN rather than QRNN | TrecQA TRAIN-ALL · TrecQA TRAIN-ALL | MRR: about $1\%-5\%$ better than QRNN | Helped |
| Use index-wise gate alignment instead of soft-attention alignment | CTRN lightweight temporal crossing · Question-answer ranking | performance | No clear effect |
| Cross-apply temporal gates between question and answer sequences | QRNN · CTRN parameter count | parameters: 1.05M | No clear effect |
| Use CTRN rather than QRNN | TrecQA TRAIN-ALL · TRAIN-ALL runtime | runtime per epoch: approximately $10s$ per epoch | Hurt |
Closest papers by meaning
- Focused Hierarchical RNNs for Conditional Sequence Processing (ICML 2018 · similarity 0.87)
- Ask Me Anything: Dynamic Memory Networks for Natural Language Processing (ICML 2016 · similarity 0.86)
- Recurrently Controlled Recurrent Networks (NeurIPS 2018 · similarity 0.86)
- Kernel-Based Approaches for Sequence Modeling: Connections to Neural Methods (NeurIPS 2019 · similarity 0.86)
- Spike-inspired rank coding for fast and accurate recurrent neural networks (ICLR 2022 · similarity 0.86)
- Titans: Learning to Memorize at Test Time (NeurIPS 2025 · 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.