Reproduce “Mixture of Low Rank Adaptation with Partial Parameter Sharing for Time Series Forecasting” 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
Licheng Pan, Zhichao Chen, Haoxuan Li, Guangyi Liu, Zhijian Xu, Zhaoran Liu, Hao Wang, Ying Wei
MoLA adapts forecast steps with shared low-rank experts, addressing a limitation in multi-task time-series forecasting: one linear decoder maps shared representations to every future step. It divides the horizon into segments and learns segment-specific mixtures of shared adaptations. On Weather at horizon 96, MSE is 0.173 versus 0.201 for the iTransformer MT-F baseline, a result limited to one dataset and horizon.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.