Reproduce “What Is the Influence of Task Order in Continual Learning? An Empirical Study and Order-Aware Evaluation for Motion 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
Zirui Li, Yunlong Lin, Haohan Yang, Xiaocong Zhao, Chen Lv
The authors evaluate how task order affects continual motion forecasting across four INTERACTION road-layout tasks, using 24 orders, four algorithms, and three training replicates per order–algorithm pairing. They decompose variation across orders, algorithms, interactions, and training runs. Order explains 64.6%–91.3% of variation in four error metrics, chiefly through starting-task identity; this finding is specific to this benchmark.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.