Research map / Architectures and efficiency / RNNs and normalization
State-Frequency Memory 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 |
|---|---|---|---|
| Adding adaptive (input-dependent) frequencies instead of static frequencies (ablation) | SFM · TIMIT phoneme classification | frame accuracy | Helped |
| Using static frequencies (SFM) vs adaptive frequencies (ablation) | SFM · Polyphonic music (JSB chorales, Piano-midi.de, MuseData) | log-likelihood | No clear effect |
| SFM frequency decomposition vs classic RNN/LSTM baselines | SFM · Polyphonic music (MuseData, JSB chorales, Piano-midi.de) | log-likelihood: about 0.2 to 1.0 better than BG1 and BG2 baselines | Helped |
| SFM frequency decomposition vs RNN-type models on Nottingham | SFM · Nottingham polyphonic music | log-likelihood | Hurt |
| SFM frequency decomposition vs LSTM baselines on phoneme classification | SFM · TIMIT phoneme classification | frame accuracy: more than 5% over LSTM, CW-RNN and ACT-RNN | Helped |
| Adaptive SFM vs Bi-LSTM on phoneme classification | Adaptive SFM · TIMIT phoneme classification | frame accuracy: slightly better | Mixed |
Closest papers by meaning
- Combining Recurrent, Convolutional, and Continuous-time Models with Linear State Space Layers (NeurIPS 2021 · similarity 0.88)
- Understanding and Controlling Memory in Recurrent Neural Networks (ICML 2019 · similarity 0.88)
- Phased LSTM: Accelerating Recurrent Network Training for Long or Event-based Sequences (NeurIPS 2016 · similarity 0.88)
- Recurrent neural networks: vanishing and exploding gradients are not the end of the story (NeurIPS 2024 · similarity 0.87)
- Fractional State Space Transition for Long Sequence Modeling (NeurIPS 2026 · similarity 0.86)
- Learning dynamics in linear recurrent neural networks (ICML 2025 · similarity 0.86)
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.