Research map / Time series / Foundation models for time series
Learning filter widths of spectral decompositions with wavelets
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 WD layer in place of mel-filter bank preprocessing | CNN · TIMIT | phone error rate (%): 18.9 -> 18.1 | Helped |
| add WD layer as first layer of CNN | CNN · UCR Haptics | test error: 0.449 -> 0.425 | Helped |
| replace first CNN layer with WD layer (same parameter count) | 4-layer CNN · synthetic frequency-order dataset | AUC-ROC / learning speed | Helped |
| WD+CNN versus the best hierarchical CNN with engineered features | WD + CNN · TIMIT | phone error rate (%): 16.5 vs 18.1 | Hurt |
| use Adam instead of SGD with fixed learning rate for scale learning (ablation) | WD layer · training speed | convergence speed | Helped |
Closest papers by meaning
- A General Graph Spectral Wavelet Convolution via Chebyshev Order Decomposition (ICML 2025 · similarity 0.85)
- Content-aware Balanced Spectrum Encoding in Masked Modeling for Time Series Classification (AAAI 2025 · similarity 0.84)
- Time Series Representations with Hard-Coded Invariances (ICML 2025 · similarity 0.84)
- FilterNet: Harnessing Frequency Filters for Time Series Forecasting (NeurIPS 2024 · similarity 0.83)
- MPTSNet: Integrating Multiscale Periodic Local Patterns and Global Dependencies for Multivariate Time Series Classification (AAAI 2025 · similarity 0.83)
- Spectral Transformer Neural Processes (NeurIPS 2026 · similarity 0.83)
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.