中文

真实场景下的量化(Quantification):数据集与基线

机器学习 2015-12-01 v2

摘要

量化(Quantification)是估计数据集类别分布的任务。虽然通常被视为对数据集偏移有严格假设的参数估计问题,我们考虑真实场景下的量化(Quantification),基于两个来自海洋生态学的大规模数据集:加勒比珊瑚礁调查和玛莎葡萄园海岸观测站的浮游生物时间序列。我们调研了文献中的几种量化方法,并指出未来工作的机会。特别地,我们展示了一个深度神经网络可以在极少量数据(25-100个样本)上微调,以超越替代方法。

关键词

引用

@article{arxiv.1510.04811,
  title  = {Quantification in-the-wild: data-sets and baselines},
  author = {Oscar Beijbom and Judy Hoffman and Evan Yao and Trevor Darrell and Alberto Rodriguez-Ramirez and Manuel Gonzalez-Rivero and Ove Hoegh - Guldberg},
  journal= {arXiv preprint arXiv:1510.04811},
  year   = {2015}
}

备注

This report was prsented at the NIPS 2015 workshop on Transfer and Multi-Task Learning: Trends and New Perspectives. It is 4 pages + 1 page of references followed by a 6 page appendix