Bradley-Terry模型的在线密度估计Online Density Estimation of Bradley-Terry Models |
|
课程网址: | https://videolectures.net/videos/colt2015_takimoto_online_density |
主讲教师: | Eiji Takimoto |
开课单位: | 信息不详。欢迎您在右侧留言补充。 |
开课时间: | 2025-02-04 |
课程语种: | 英语 |
中文简介: | 我们考虑了Bradley-Terry模型的在线密度估计问题,该问题确定了$n$组中任意对之间匹配结果的概率。一个恼人的问题是损失函数不是凸的。避免非凸性的一个标准解决方案是改变变量,使新的损失函数w.r.t.新变量是凸的。但是,新定义域的半径$K$可能很大,或者一般来说是无穷大。当$K$为常数时,标准算法OGD和ONS分别有遗憾界$O(n^{\frac{1}{2}}(\ln K)\sqrt{T})$和$O(n^{\frac{3}{2}}K\ln T)$。 |
课程简介: | We consider an online density estimation problem for Bradley-Terry models which determine the probability of a match result between any pair in the set of $n$ teams. An annoying issue is that the loss function is not convex. A standard solution to avoid the non-convexity is to change variables so that the new loss function w.r.t.~new variables is convex. But, then the radius $K$ of the new domain might be huge or infinite in general. When $K$ is regarded as a constant, standard algorithms OGD and ONS have regret bounds $O(n^{\frac{1}{2}}(\ln K)\sqrt{T})$ and $O(n^{\frac{3}{2}}K\ln T)$, respectively. |
关 键 词: | 密度估计问题; 损失函数; 非凹性 |
课程来源: | 视频讲座网 |
数据采集: | 2025-03-28:zsp |
最后编审: | 2025-03-28:zsp |
阅读次数: | 4 |