摘要: (转载自:WikiPedia) Fine tuning is a process to take a network model that has already been trained for a given task, and make it perform a second similar 阅读全文
posted @ 2018-11-29 18:32 leizhao 阅读(366) 评论(0) 推荐(0) 编辑
摘要: by Jason Brownlee on December 20, 2017 in Better Deep Learning Transfer learning is a machine learning method where a model developed for a task is re 阅读全文
posted @ 2018-11-29 17:24 leizhao 阅读(538) 评论(0) 推荐(0) 编辑
摘要: 我的理解: 经验风险,基于样本学习的经验进行决策的错误风险(或称由样本学习的经验进行决策所造成的损失)。将经验风险最小化,将会使模型对训练样本的学习能力增强,表现为拟合能力增强。显然,一味的经验风险最小化,会使得模型对训练样本过拟合。 期望风险,所谓期望,是对所有可能输入的分布的预测,假定输入的分布 阅读全文
posted @ 2018-11-29 17:04 leizhao 阅读(1353) 评论(0) 推荐(0) 编辑