The Power of Subsampling in Submodular Maximization

被引:3
作者
Harshaw, Christopher [1 ]
Kazemi, Ehsan [2 ]
Feldman, Moran [3 ]
Karbasi, Amin [1 ,4 ,5 ]
机构
[1] Yale Univ, Dept Comp Sci, New Haven, CT 06520 USA
[2] Google, CH-8002 Zurich, Switzerland
[3] Univ Haifa, Dept Comp Sci, IL-3498838 Haifa, Israel
[4] Yale Univ, Dept Elect Engn, New Haven, CT 06520 USA
[5] Yale Univ, Dept Stat & Data Sci, New Haven, CT 06520 USA
基金
美国国家科学基金会; 以色列科学基金会;
关键词
submodular maximization; subsampling; streaming algorithms; approximation algorithms; p-extendible systems; p-matchoids; APPROXIMATIONS; ALGORITHM; GREEDY;
D O I
10.1287/moor.2021.1172
中图分类号
C93 [管理学]; O22 [运筹学];
学科分类号
070105 ; 12 ; 1201 ; 1202 ; 120202 ;
摘要
We propose subsampling as a unified algorithmic technique for submodular maximization in centralized and online settings. The idea is simple: independently sample elements from the ground set and use simple combinatorial techniques (such as greedy or local search) on these sampled elements. We show that this approach leads to optimal/state-of-the-art results despite being much simpler than existing methods. In the usual off-line setting, we present SAMPLEGREEDY, which obtains a (p + 2 + o(1))-approximation for maximizing a submodular function subject to a p-extendible system using O(n + nk/p) evaluation and feasibility queries, where k is the size of the largest feasible set. The approximation ratio improves to p + 1 and p for monotone submodular and linear objectives, respectively. In the streaming setting, we present Sample-Streaming, which obtains a (4p + 2 - o(1))-approximation for maximizing a submodular function subject to a p-matchoid using O(k) memory and O(km/p) evaluation and feasibility queries per element, and m is the number of matroids defining the p-matchoid. The approximation ratio improves to 4p for monotone submodular objectives. We empirically demonstrate the effectiveness of our algorithms on video summarization, location summarization, and movie recommendation tasks.
引用
收藏
页码:1365 / 1393
页数:29
相关论文
共 68 条
  • [1] Badanidiyuru A., 2014, P 25 ANN ACM SIAM S, P1497, DOI [DOI 10.1137/1.9781611973402.110, 10.1137/1.9781611973402. 110]
  • [2] Badanidiyuru A., 2020, ADV NEURAL INFORM PR
  • [3] Streaming Submodular Maximization: Massive Data Summarization on the Fly
    Badanidiyuru, Ashwinkumar
    Mirzasoleiman, Baharan
    Karbasi, Amin
    Krause, Andreas
    [J]. PROCEEDINGS OF THE 20TH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING (KDD'14), 2014, : 671 - 680
  • [4] Balkanski E, 2016, PR MACH LEARN RES, V48
  • [5] Buchbinder N, 2014, P 25 ANN ACM SIAM S, P1433, DOI [10.1137/1.9781611973402.106, DOI 10.1137/1.9781611973402.106]
  • [6] Buchbinder N., 2015, P 26 ANN ACM SIAM S, P1202
  • [7] Buchbinder N., 2018, SUBMODULAR FUNCTIONS
  • [8] Constrained Submodular Maximization via a Nonsymmetric Technique
    Buchbinder, Niv
    Feldman, Moran
    [J]. MATHEMATICS OF OPERATIONS RESEARCH, 2019, 44 (03) : 988 - 1005
  • [9] Online Submodular Maximization: Beating 1/2 Made Simple
    Buchbinder, Niv
    Feldman, Moran
    Filmus, Yuval
    Garg, Mohit
    [J]. INTEGER PROGRAMMING AND COMBINATORIAL OPTIMIZATION, IPCO 2019, 2019, 11480 : 101 - 114
  • [10] Comparing Apples and Oranges: Query Trade-off in Submodular Maximization
    Buchbinder, Niv
    Feldman, Moran
    Schwartz, Roy
    [J]. MATHEMATICS OF OPERATIONS RESEARCH, 2017, 42 (02) : 308 - 329