Reproduce “Deep Time-Series Forecasting in 10 Years: A Survey” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hao Wang, Licheng Pan, Qingsong Wen, Jialin Yu, Zhichao Chen, Chunyuan Zheng, Xiaoxi Li, Zhixuan Chu and 7 more
The authors survey deep time-series forecasting through autocorrelation, organizing methods around modeling dependence in both the history and the future label sequence. They classify architectures that represent past readings and losses that account for structure among future values. In a shared-backbone comparison, five correlation-aware losses beat pointwise MSE on average across four datasets, but no loss wins uniformly.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.