TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting

Key points

• However, real-world time series are often noisy and non-stationary, making deep learning models prone to fitting incidental temporal interactions in the training data.1
• We analyze temporal relationship matrices on commonly used time series datasets and find that samples from different temporal positions and different variables exhibit shared dominant temporal relationship structures.1
• To learn such shared structures, we propose Temporal Structure Network (TStruct), a simple yet effective network for directly learning shared and generalizable temporal dependencies.1
• These weights represent a soft allocation of each sample-variable pair over shared temporal dependency structures, enabling the model to generate adaptive temporal dependency matrices under shared structural constraints, rather than fitting incidental interactions in a fully unconstrained manner.1

Sources (1)

• [1]TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 07:24 AM

  Long-term time series forecasting requires temporal dependency modeling that remains stable and generalizable under temporal distribution shifts.

  However, real-world time series are often noisy and non-stationary, making deep learning models prone to fitting incidental temporal interactions in the training data.

Extractive summary: sentences quoted from the sources.