• <div id="oaooe"></div>
  • <li id="oaooe"><s id="oaooe"></s></li><li id="oaooe"><button id="oaooe"></button></li>
  • <li id="oaooe"><button id="oaooe"></button></li>
  • <div id="oaooe"><button id="oaooe"></button></div>
  • <li id="oaooe"><button id="oaooe"></button></li>
  • <li id="oaooe"></li>
  • <li id="oaooe"><button id="oaooe"></button></li>
  • <small id="oaooe"></small>
  • <div id="oaooe"><li id="oaooe"></li></div><small id="oaooe"><button id="oaooe"></button></small>
    <div id="oaooe"><s id="oaooe"></s></div>
  • 信息門戶
    ENGLISH
    舊版回顧
    當前位置: 首頁>>學術講座>>正文

    關于北京工業大學徐大川教授學術報告的通知

    發布時間:2021-04-13 作者:  來源: 閱讀量:[]

    報告題目:Outliers Detection Is Not So Hard: Approximation Algorithms for Robust Clustering Problems Using Local Search Techniques

    報告人:徐大川 教授

    報告時間:4月16日15:00-16:00

    報告地點:6號教學樓303室

    報告簡介:

    In this talk, we consider two types of robust models of the k-median/k-means problems: the outlier-version (k-MedO/k-MeaO) and the penalty-version (k-MedP /k-MeaP), in which we can mark some points as outliers and discard them. In k-MedO /k-MeaO, the number of outliers is bounded by a given integer. In k-MedO/k-MeaO, we do not bound the number of outliers, but each outlier will incur a penalty cost. We develop a new technique to analyze the approximation ratio of local search algorithms for these two problems by introducing an adapted cluster that can capture useful information about outliers in the local and the global optimal solution. For k-MeaP, we improve the best known approximation ratio based on local search from 25+\epsilon to 9+\epsilon. For k-MedP, we obtain the best known approximation ratio. For k-MedO/k-MeaO, there exists only two bi-criteria approximation algorithms based on local search. One violates the outlier constraint (the constraint on the number of outliers), while the other violates the cardinality constraint (the constraint on the number of clusters). We consider the former algorithm and improve its approximation ratios from 17+\epsilon to 3+\epsilon for k-MedO, and from 274+\epsilon to 9+\epsilon for k-MeaO. (Joint work with Yishui Wang, Rolf H. Mohring, Chenchen Wu, and Dongmei Zhang)

    報告人簡介:

    徐大川,北京工業大學數學學院運籌學與控制論責任教授,數學/統計學博士生導師。北京工業大學區塊鏈研究中心副主任。2002年于中國科學院數學與系統科學研究院獲得博士學位。研究興趣包括:組合優化、近似算法、機器學習等。中國運籌學會數學規劃分會理事長,中國運籌學會常務理事,北京運籌學會副理事長。擔任AMC、APJOR、JORSC、運籌與管理等期刊編委。在科學出版社出版學術專著《設施選址問題的近似算法》,在Mathematical Programming,Operations Research,INFORMS Journal on Computing,Omega,Algorithmica,Journal of Global Optimization,Theoretical Computer Science,Information Process Letters,Journal of Combinatorial Optimization,Operations Research Letters等期刊和AAAI, ICDCS,COCOON等會議發表學術論文100余篇。


    理學院

    2021年4月13日

    上一條:關于《構建區塊鏈互聯生態的思考與實踐》學術報告通知
    下一條:關于開展青年學者沙龍活動的通知

    關閉

    相關鏈接
     
    地址:天津市西青區賓水西道391號 郵編:300384  
    電話:0086-22-60215678  傳真:0086-22-60215555  
    本科招生電話:0086-22-60216795
    ICP備案號:津ICP備11001169號-1|津教備0064號
    杏吧直播app官方下载 百度 好搜 搜狗

    警告:本站禁止未滿18周歲訪客瀏覽,如果當地法律禁止請自覺離開本站!收藏本站:請使用Ctrl+D進行收藏