Structured Latent Space Modeling over Multi-Scale Temporal Patches for Multivariate Time Series Forecasting

Name: Structured Latent Space Modeling over Multi-Scale Temporal Patches for Multivariate Time Series Forecasting
Uploaded: 2026-09-05T06:28:40.993715+00:00
Duration: 425 s
Description: This episode discusses M2Patch, a method for multivariate time series forecasting. The model uses multi-scale temporal patches to capture signals moving at different speeds and builds a structured latent space. By employing depthwise separable convolutions, it achieves linear computational complexity and strong performance across various real-world benchmarks.

This episode discusses M2Patch, a method for multivariate time series forecasting. The model uses multi-scale temporal patches to capture signals moving at different speeds and builds a structured latent space. By employing depthwise separable convolutions, it achieves linear computational complexity and strong performance across various real-world benchmarks.

Open the full episode