CS Peer Talk | Simple & Optimal Quantile Sketch…

215次阅读
没有评论

CS Peer Talk | Simple & Optimal Quantile Sketch...

第三十四期

报告人:武弘勋,UC Berkeley

时间:12月20日(星期三)4:00pm

地点:静园五院204

报告信息

Title

Simple & Optimal Quantile Sketch—Combining Greenwald-Khanna with Khanna-Greenwald

Abstract

Finding the approximate median (or more generally, quantiles) in a stream is a fundamental problem in data science. 20 years ago, Greenwald and Khanna proposed a deterministic algorithm for approximate streaming quantiles, whose space complexity was later proven to be optimal. This algorithm is very popular and gets “implemented” in spark-SQL and other software packages.

However, since it is complicated, these “implementations” are all oversimplified heuristics without theoretical guarantees. In this talk, I will present an algorithm that achieves both simplicity and optimality.

Biography

 CS Peer Talk | Simple & Optimal Quantile Sketch...

Hongxun Wu is a second-year PhD student at the theory group of UC Berkeley, advised by Jelani Nelson and Avishay Tal.His research interest is in low space computation, especially its derandomization and streaming algorithms.

about CS Peer Talk

作为活动的发起人,我们来自北京大学图灵班科研活动委员会,主要由图灵班各年级同学组成。我们希望搭建一个CS同学交流的平台,促进同学间的交流合作,帮助同学练习展示,同时增进友谊。

目前在计划中的系列包括但不限于:

  • 教程系列学生讲者为主,介绍自己的研究领域

  • 研究系列学生讲者为主,介绍自己的研究成果

  • 客座系列邀请老师做主题报告

除非报告人特别要求,报告默认是非公开的,希望营造一个自由放松但又互相激励的交流氛围。

CS Peer Talk | Simple & Optimal Quantile Sketch...

主讲嘉宾招募

如果你愿意和大家分享你的学术成果、经历经验,总结回顾、触发新思,欢迎报名自荐


主讲人报名:发邮件至 cs_research_tc@163.com,写明想讲的题目、内容及时间。

CS Peer Talk | Simple & Optimal Quantile Sketch...

CS Peer Talk | Simple & Optimal Quantile Sketch...

北京大学图灵班科研活动委员会

CS Peer Talk | Simple & Optimal Quantile Sketch...

—   版权声明  —

本微信公众号所有内容,由北京大学前沿计算研究中心微信自身创作、收集的文字、图片和音视频资料,版权属北京大学前沿计算研究中心微信所有;从公开渠道收集、整理及授权转载的文字、图片和音视频资料,版权属原作者。本公众号内容原作者如不愿意在本号刊登内容,请及时通知本号,予以删除。

CS Peer Talk | Simple & Optimal Quantile Sketch...

 

Read More 

正文完
可以使用微信扫码关注公众号(ID:xzluomor)
post-qrcode
 
评论(没有评论)
Generated by Feedzy