Research map / Architectures and efficiency / RNNs and normalization
Structured Sparsification of Gated Recurrent Neural Networks
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 gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · IMDb | compression: 17874x -> 19747x | Helped |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · AGNews | compression: 645x -> 647x | Helped |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · AGNews | accuracy (%): 88.55 -> 88.41 | Hurt |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · Char PTB | compression: 10.2x -> 9.8x | Hurt |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · Char PTB | bits-per-char: 1.478 - 1.430 -> 1.467 - 1.425 | Helped |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · Word PTB small | perplexity: 110.25 -- 104.81 -> 109.98 -- 104.45 | Helped |
| Add gate sparsification to weight and neuron sparsification (ablation) | Bayes W+N · Word PTB small | compression: 11.65x -> 11.44x | Hurt |
| Add gate sparsification to weight and neuron pruning (ablation) | Prun. W+N · Word PTB small | perplexity: 110.34 -- 106.25 -> 110.04 -- 105.64 | Helped |
| Add gate sparsification to weight and neuron pruning (ablation) | Prun. W+N · Word PTB small | compression: 1.44x -> 1.49x | Helped |
| Add gate sparsification to weight and neuron pruning (ablation) | Prun. W+N · Word PTB large | compression: 2.97x -> 3.22x | Helped |
| Add gate sparsification to weight and neuron pruning (ablation) | Prun. W+N · Word PTB large | perplexity: 81.25 -- 77.62 -> 81.24 -- 77.82 | Mixed |
| Use unstructured weight sparsification without gate or neuron group weights (ablation) | Bayes W+G+N · IMDb | compression: 19747x -> 18567 | Hurt |
Closest papers by meaning
- Learning Intrinsic Sparse Structures within Long Short-Term Memory (ICLR 2018 · similarity 0.90)
- From Bayesian Sparsity to Gated Recurrent Nets (NeurIPS 2017 · similarity 0.89)
- Towards Binary-Valued Gates for Robust LSTM Training (ICML 2018 · similarity 0.88)
- Sparsing Law: Towards Large Language Models with Greater Activation Sparsity (ICML 2025 · similarity 0.86)
- Accelerating Linear Recurrent Neural Networks for the Edge with Unstructured Sparsity (ICML 2025 · similarity 0.86)
- The Sparsity Whisperer (NeurIPS 2026 · 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.