|
|
报告题目: 大数据下分布计算(Distributed subset selection)
时间: 6月7号上午10点
地点: 伊泰利大厦2号楼301
报告人: 焦雨领, 副教授
中南财经政法大学统计与数学学院统计学系,文澜学者。主要从事统计计算,机器学习,反问题,图像处理,数据科学中的建模和随机、并行、分布式算法的研究。主持国家自然科学基金青年项目和湖北省自然科学基金面上项目各一项,中央高校研究项目3项。
In this talk I will present a distributed homotopy generalized Newton (HotGeNew) algorithm for computing the subset selection problem in high dimension. Under certain regularity condition, HotGeNew will get the oracle estimator in finite steps. We also consider a distributed implementation of HotGeNew when the data is distributed stored. We will show the distributed HotGeNew output an estimator with minimax estimation error in a controllable number of rounds of communications.
|
|
|
|