arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1902.06789cs.LGstat.ML
收藏

机器学习研究中的七个神话

Seven Myths in Machine Learning Research

Oscar Chang, Hod Lipson

摘要:

我们提出了机器学习研究中普遍认为正确的七个神话,大约在2019年2月。这是www.example.com上博客文章的存档副本https://crazyoscarchang.github.io/2019/02/16/seven-myths-in-machine-learning-research/神话1:误解2:图像数据集代表在野外发现的真实图像误解3:机器学习研究人员不使用测试集进行验证误解4:每个数据点都用于训练神经网络误解5:我们需要(批量)归一化来训练非常深的残差网络误解6:注意$>$卷积误解7:显着图是解释神经网络的可靠方法

英文摘要:

We present seven myths commonly believed to be true in machine learning research, circa Feb 2019. This is an archival copy of the blog post at https://crazyoscarchang.github.io/2019/02/16/seven-myths-in-machine-learning-research/ Myth 1: TensorFlow is a Tensor manipulation library Myth 2: Image datasets are representative of real images found in the wild Myth 3: Machine Learning researchers do not use the test set for validation Myth 4: Every datapoint is used in training a neural network Myth 5: We need (batch) normalization to train very deep residual networks Myth 6: Attention $>$ Convolution Myth 7: Saliency maps are robust ways to interpret neural networks

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑