Research map / Architectures and efficiency / RNNs and normalization
Knowledge-aware Leap-LSTM: Integrating Prior Knowledge into Leap-LSTM towards Faster Long Text Classification
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 knowledge-oriented cost Lk to Factored KALL (ablation) | Factored KALL · Full Small | Micro F1: 55.73 -> 56.50 | Helped |
| Add the knowledge-oriented cost Lk to Factored KALL (ablation) | Factored KALL · Polarity Small | Micro F1: 90.69 -> 91.56 | Helped |
| Add the knowledge-oriented cost Lk to Gated KALL (ablation) | Gated KALL · Full Small | Micro F1: 53.25 -> 57.48 | Helped |
| Add the knowledge-oriented cost Lk to Gated KALL (ablation) | Gated KALL · Polarity Small | Micro F1: 89.03 -> 92.71 | Helped |
| Add keyword features as factors, without Lk (ablation) | Leap-LSTM · Full Small | Micro F1: 54.08 -> 55.73 | Helped |
| Add keyword features as factors, without Lk (ablation) | Leap-LSTM · Polarity Small | Micro F1: 89.06 -> 90.69 | Helped |
| Enforce keyword inclusion with a gate, without Lk (ablation) | Leap-LSTM · Full Small at skip rate 0.6 or 0.9 | Micro F1: 59.48 -> 58.90; 54.08 -> 53.25 | Hurt |
| Add the knowledge-oriented cost Lk to Gated KALL (ablation) | Gated KALL · Full Small at skip rate 0.25 | Micro F1: 61.67 -> 62.33 | Helped |
| Add the knowledge-oriented cost Lk to Gated KALL (ablation) | Gated KALL · Polarity Small at skip rate 0.25 | Micro F1: 94.73 -> 95.02 | Helped |
| Use Gated KALL + Lk instead of XLNet (256) | Long-text classification · Full Small CPU-only inference | speedup: 23.6x faster | Helped |
Closest papers by meaning
- Neural Speed Reading with Structural-Jump-LSTM (ICLR 2019 · similarity 0.86)
- SkipW: Resource Adaptable RNN with Strict Upper Computational Limit (ICLR 2021 · similarity 0.86)
- Regularizing and Optimizing LSTM Language Models (ICLR 2018 · similarity 0.86)
- Automatic Rule Extraction from Long Short Term Memory Networks (ICLR 2017 · similarity 0.86)
- Long Short-Term Memory with Dynamic Skip Connections (AAAI 2019 · similarity 0.85)
- Extended LSTMs for Knowledge Tracing: Peeking Inside the Black Box (Student Abstract) (AAAI 2025 · similarity 0.85)
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.