CP Loss: Channel-wise Perceptual Loss for Time Series Forecasting

TL;DR

This paper introduces CP Loss, a novel channel-wise perceptual loss function for time series forecasting that captures channel-specific dynamics by learning dedicated perceptual spaces, improving model performance.

Contribution

It proposes a learnable channel-wise filter to create perceptual spaces for each channel, jointly optimized with forecasting models, to better capture heterogeneity in multi-channel time series data.

Findings

Enhanced forecasting accuracy on multi-channel data

Effective modeling of channel-specific fluctuations

Joint optimization improves perceptual space learning

Abstract

Multi-channel time-series data, prevalent across diverse applications, is characterized by significant heterogeneity in its different channels. However, existing forecasting models are typically guided by channel-agnostic loss functions like MSE, which apply a uniform metric across all channels. This often leads to fail to capture channel-specific dynamics such as sharp fluctuations or trend shifts. To address this, we propose a Channel-wise Perceptual Loss (CP Loss). Its core idea is to learn a unique perceptual space for each channel that is adapted to its characteristics, and to compute the loss within this space. Specifically, we first design a learnable channel-wise filter that decomposes the raw signal into disentangled multi-scale representations, which form the basis of our perceptual space. Crucially, the filter is optimized jointly with the main forecasting model, ensuring that the learned perceptual space is explicitly oriented towards the prediction task. Finally, losses are calculated within these perception spaces to optimize the model. Code is available at https://github.com/zyh16143998882/CP_Loss.