Research map / Architectures and efficiency / RNNs and normalization
Neural Speed Reading with Structural-Jump-LSTM
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 skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · IMDB | accuracy: 0.000 | No clear effect |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · DBPedia | accuracy: 0.013 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · Yelp | accuracy: 0.003 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · AG news | accuracy: 0.003 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · SST | accuracy: 0.004 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · Rotten Tomatoes | accuracy: 0.003 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · CBT-CN | accuracy: 0.007 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · CBT-NE | accuracy: 0.010 | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · IMDB | FLOP reduction: 6.3x | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · DBPedia | FLOP reduction: 7.0x | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · Yelp | FLOP reduction: 1.9x | Helped |
| Use skip actions and punctuation-structured jumps instead of full sequential reading | LSTM · AG news | FLOP reduction: 2.4x | Helped |
Closest papers by meaning
- Neural Speed Reading via Skim-RNN (ICLR 2018 · similarity 0.88)
- Knowledge-aware Leap-LSTM: Integrating Prior Knowledge into Leap-LSTM towards Faster Long Text Classification (AAAI 2021 · similarity 0.86)
- Fast-Slow Recurrent Neural Networks (NeurIPS 2017 · similarity 0.86)
- Mogrifier LSTM (ICLR 2020 · similarity 0.85)
- Structured in Space, Randomized in Time: Leveraging Dropout in RNNs for Efficient Training (NeurIPS 2021 · similarity 0.85)
- Long Short-Term Memory with Dynamic Skip Connections (AAAI 2019 · 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.