deeptime专题

论文笔记 Shape and Time Distortion Loss for Training DeepTime Series Forecasting Models

该论文的poster DILATE/poster_DILATE.pdf at master · vincent-leguen/DILATE (github.com) 0 摘要         本文解决了非平稳信号的时间序列预测和多个未来步骤预测的问题。为了处理这个具有挑战性的任务,我们引入了 DILATE(包括形状和时间的失真损失,DIstortion Loss including shAp