Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Volatility Based Kernels and Moving Average Means for Accurate Forecasting with Gaussian Processes
Gregory Benton, Wesley Maddox, Andrew Gordon Wilson
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 the Volt volatility kernel instead of a Matérn kernel | GP forecasting · Stock Prices | NLL: 7.74 \pm 0.21 -> 4.69 \pm 0.03 | Improved |
| Use the Volt volatility kernel instead of a Matérn kernel | GP forecasting · Wind Speeds | NLL: 18.03 \pm 1.90 -> 3.38 \pm 0.05 | Improved |
| Use the Volt volatility kernel instead of a Matérn kernel | GP forecasting · FX | NLL: -0.36 \pm 0.04 -> -1.60 \pm 0.02 | Improved |
| Add a Magpie moving-average mean (ablation) | Volt · Stock Prices | NLL: 4.69 \pm 0.03 -> 5.88 \pm 0.02 | Hurt |
| Add a Magpie moving-average mean (ablation) | Volt · Wind Speeds | NLL: 3.38 \pm 0.05 -> 4.28 \pm 0.16 | Hurt |
| Add a Magpie moving-average mean (ablation) | Volt · FX | NLL: -1.60 \pm 0.02 -> -1.69 \pm 0.02 | Improved |
| Add a Magpie mean instead of a constant mean (ablation) | Volt · Stock price forecasts | calibration | Improved |
| Add a Magpie mean instead of a constant mean (ablation) | Volt · Wind speed forecasts | calibration | Improved |
| Use Volt rather than standard GP kernels or probabilistic LSTMs | Probabilistic forecasting · NASDAQ 100 stock prices | calibration | Improved |
| Add a small amount of mean reversion to rollout forecasts | Volt · Wind Speeds | calibration | Improved |
| Use multitask Volt rather than independent Volt | Volt · Grouped stock forecasting | calibration error | Improved |
Closest papers by meaning
- From GARCH to Neural Network for Volatility Forecast (AAAI 2024 · similarity 0.84)
- Probabilistic Forecasting with Stochastic Interpolants and Föllmer Processes (ICML 2024 · similarity 0.84)
- Robust and Conjugate Spatio-Temporal Gaussian Processes (ICML 2025 · similarity 0.83)
- KernelMatmul: Scaling Gaussian Processes to Large Time Series (AAAI 2025 · similarity 0.83)
- Conditioning Gaussian Processes on Almost Anything (NeurIPS 2026 · similarity 0.83)
- A Set-Sequence Model for Time Series (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.
