首页>
外国专利>
Device for transfer learning between modified tasks
Device for transfer learning between modified tasks
展开▼
机译:用于在修改后的任务之间进行转移学习的设备
展开▼
页面导航
摘要
著录项
相似文献
摘要
Device for transfer learning of hyperparameters of a machine learning algorithm, the device comprising a machine-readable storage medium on which instructions are stored which, when executed by a computer, caused the computer to execute a method comprising the following steps: i) Provision of a current search space (X i ) and a previous search space (X i-1 ), a cost function (f) to be optimized, and a ranking (C i-1 ) of evaluated hyperparametric configurations of a previous optimization step using the previous search space (X i-1 ) with regard to the cost function (f), wherein the search spaces (X i, X i-1 ) are each defined on the basis of predetermined value ranges of the hyperparameters; i) Creating a reduced search area ( ) where the value ranges of the hyperparameters of the reduced search space ( ) corresponds to the value ranges of the hyperparameters of the current search space (X i ), restrictedly dependent on the values of a predeterminable number (a) of the best hyperparameter configurations from the ranking (C i-1 ); iii) Multiple, random pulling of candidate configurations from the reduced search space ( ) and the current search space X i and applying the machine learning algorithm, each parameterized with the candidate configurations for optimizing the cost function (f); iv) creating a Tree Parzen Estimator (TPE) depending on the candidate solutions and the results of the machine learning algorithm applied to the cost function (f) to be optimized; v) repeating a drawing of further candidate configurations by means of the TPE from the current search space (X i ) and applying the machine learning algorithm parameterized with the candidate configurations to the cost function; vi) creating a new ranking (C i ) of the further candidate configurations and selecting a configuration from the new ranking (C i ) as a hyperparametric configuration for the machine learning algorithm.
展开▼
机译:用于转移学习机器学习算法的超参数的设备,该设备包括机器可读存储介质,在该机器可读存储介质上存储有指令,当计算机执行该指令时,该指令使计算机执行包括以下步骤的方法:i)提供当前搜索空间(X i Sub>)和先前搜索空间(X i-1 Sub>),要优化的成本函数(f)和排名(C <关于成本函数(f)的使用先前搜索空间(X i-1 Sub>)的先前优化步骤的评估超参数配置的Sub> i-1 Sub>),其中搜索空格(X i, Sub> X i-1 Sub>)是根据超参数的预定值范围定义的; i)创建一个缩小的搜索区域(),其中缩小的搜索空间()的超参数的值范围与当前搜索空间(X i Sub>)的超参数的值范围相对应根据排名(C i-1 Sub>)的最佳超参数配置的可确定数量(a)的值; iii)从缩小的搜索空间()和当前搜索空间X i Sub>中随机抽取候选配置,并应用机器学习算法,每项参数均通过候选配置进行参数化,以优化成本函数(f ); iv)根据候选解决方案和应用于待优化成本函数(f)的机器学习算法的结果,创建一个Tree Parzen估计器(TPE); v)通过TPE从当前搜索空间(X i Sub>)重复绘制其他候选配置的图,并将用候选配置参数化的机器学习算法应用于成本函数; vi)创建其他候选配置的新排名(C i Sub>),并从新排名(C i Sub>)中选择一个配置作为机器学习算法的超参数配置。
展开▼