CASA: CNN Autoencoder-based Score Attention for Efficient Multivariate Long-term Time-series Forecasting
This work addresses efficiency and performance issues in time-series forecasting for applications like weather prediction and traffic analysis, representing a novel method for a known bottleneck.
The paper tackles the limitations of time complexity, computational resources, and cross-dimensional interactions in multivariate long-term time-series forecasting by introducing CASA, a CNN Autoencoder-based Score Attention mechanism, which reduces computational resources by up to 77.7%, accelerates inference by 44.0%, and achieves state-of-the-art performance in 87.5% of evaluated metrics.
Multivariate long-term time series forecasting is critical for applications such as weather prediction, and traffic analysis. In addition, the implementation of Transformer variants has improved prediction accuracy. Following these variants, different input data process approaches also enhanced the field, such as tokenization techniques including point-wise, channel-wise, and patch-wise tokenization. However, previous studies still have limitations in time complexity, computational resources, and cross-dimensional interactions. To address these limitations, we introduce a novel CNN Autoencoder-based Score Attention mechanism (CASA), which can be introduced in diverse Transformers model-agnosticically by reducing memory and leading to improvement in model performance. Experiments on eight real-world datasets validate that CASA decreases computational resources by up to 77.7%, accelerates inference by 44.0%, and achieves state-of-the-art performance, ranking first in 87.5% of evaluated metrics.