GO-MTL:Learning and overlap in Multi-task Learning 论文随笔
2016-12-09 10:19
399 查看
相关名词:
trace—norm constraint:
sparse matrix:就是含零项的矩数越多,越稀疏阵,零的个数越多,越稀疏
mixed integer programming :混合整数规划
nonparametric Bayesian approach:???
task clusters:??
the difference between cluster and goup:??
Newton-Raphson:??
iterative reweighted least squares algorithm(IRLS):??
normal distribution:正太分布
l1 norm : trace-norm
root mean square error: RMSE(均方根误差)
前提假设分类:
所有tasks相关
组内相关,其他的任务与别的都不相关(outlier task)
tasks 参数在低维空间,特征只有两种情况:要么对所有的任务都活跃,要么对大部分的任务都不活跃
在子空间的分组结构基于正则化,每个组内的tasks对于特征提取都共享一个线性变换 (???)
指定任务的预测性结构有L矩阵决定,而分组结构取决于s矩阵
Kronecher product:
牛顿-拉夫森(Newton-Raphson)迭代法:
Amijo rule:
假设的特点:
假设每一个任务组的参数都在一个维度的子空间,即不假设不相交任务结构,过着任务属于不同的组其他组的任务相交,通过共享一个或者多个隐藏任务
trace—norm constraint:
sparse matrix:就是含零项的矩数越多,越稀疏阵,零的个数越多,越稀疏
mixed integer programming :混合整数规划
nonparametric Bayesian approach:???
task clusters:??
the difference between cluster and goup:??
Newton-Raphson:??
iterative reweighted least squares algorithm(IRLS):??
normal distribution:正太分布
l1 norm : trace-norm
root mean square error: RMSE(均方根误差)
前提假设分类:
所有tasks相关
组内相关,其他的任务与别的都不相关(outlier task)
tasks 参数在低维空间,特征只有两种情况:要么对所有的任务都活跃,要么对大部分的任务都不活跃
在子空间的分组结构基于正则化,每个组内的tasks对于特征提取都共享一个线性变换 (???)
指定任务的预测性结构有L矩阵决定,而分组结构取决于s矩阵
Kronecher product:
牛顿-拉夫森(Newton-Raphson)迭代法:
Amijo rule:
假设的特点:
假设每一个任务组的参数都在一个维度的子空间,即不假设不相交任务结构,过着任务属于不同的组其他组的任务相交,通过共享一个或者多个隐藏任务
相关文章推荐
- Learning Task Grouping and Overlap in Multi-Task Learning
- 论文复现报告:Deep Region and Multi-label Learning for Facial Action Unit Detection
- 【CV论文阅读】Detecting events and key actors in multi-person videos
- Deep Region and Multi-label Learning for Facial Action Unit Detection简要论文笔记
- 论文笔记:Research and Implementation of a Multi-label Learning Algorithm for Chinese Text Classification
- 【论文阅读笔记】Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique
- sed and awk process xml file in multi line
- Some notes in learning MVC and EF
- A meta-model for the analysis and design of organizations in multi-agent systems
- A model for the structural, functional, and deontic specification of organizations in multiagent systems
- The mathematics in computer vision and machine learning
- [论文笔记] Money, glory and cheap talk: analyzing strategic behavior of contestants in simultaneous crowdsourcing contests on TopCoder.com (WWW, 2010)
- Graham King » In-memory key-value store in C, Go and Python
- 32-bit Assembler is Easy, why and how to develop using the assembler; start learning to program in Assembly now!
- 论文读书笔记-topic correlation and individual influenceanalysis in online forums
- 读论文笔记:Unsupervised Joint Object Discovery and Segmentation in Internet Images
- Variable OR State Visiblity:Stale Data, Locking and Visiblity and Volatile(So many Unnoticed aspect in MultiThread)
- A guide to using Paneled Forms, Multi-Splash Screens, SQL Express, and many more in Windows Application Development
- Description of CWnd derived MFC objects and multithreaded applications in Visual C++
- How to bind and validate in MultiActionController