In this work, we tackle two widespread challenges in real applications for time-series forecasting that have been largely understudied: distribution shifts and missing data. We propose SpectraNet, a novel multivariate time-series forecasting model that dynamically infers a latent space spectral decomposition to capture current temporal dynamics and correlations on the recent observed history. A Convolution Neural Network maps the learned representation by sequentially mixing its components and refining the output. Our proposed approach can simultaneously produce forecasts and interpolate past observations and can, therefore, greatly simplify production systems by unifying imputation and forecasting tasks into a single model. SpectraNet achieves SoTA performance simultaneously on both tasks on five benchmark datasets, compared to forecasting and imputation models, with up to 92% fewer parameters and comparable training times. On settings with up to 80% missing data, SpectraNet has average performance improvements of almost 50% over the second-best alternative. Our code is available at https://github.com/cchallu/spectranet.
翻译:在这项工作中,我们应对了在时间序列预测实际应用方面普遍存在的两大挑战,这些应用在很大程度上没有得到充分研究:分布变化和缺失的数据。我们提议SpectraNet,这是一个全新的多变时间序列预测模型,动态地推断出潜在的空间光谱分解,以捕捉当前时间动态和与最近观察到的历史相关关系。一个革命神经网络通过按顺序混合其组成部分和完善产出来绘制所学表现图。我们提议的方法可以同时产生预测和对过去观测进行内插,从而通过将估算和预测任务统一成单一模型,大大简化生产系统。SpectraNet在五个基准数据集的两个任务上都同时实现了 SoTA的绩效,而预测和估算模型则减少了92%的参数和可比的培训时间。在高达80%的数据缺失的情况下,SpectraNet在第二最佳替代品上平均提高了近50%的性能。我们的代码可以在https://github.com/cchallu/spectranet上查阅。