Using Temperature Sampling to Effectively Train Robot Learning Policies on Imbalanced Datasets figure
AlphaXiv 中文论文页面(可滚动查看)