On the Approximation Relationship between Optimizing Ratio of Submodular (RS) and Difference of Submodular (DS) Functions

We demonstrate that from an algorithm guaranteeing an approximation factor for the ratio of submodular (RS) optimization problem, we can build another algorithm having a different kind of approximation guarantee — weaker than the classical one — for the difference of submodular (DS) optimization problem, and vice versa. We also illustrate the link between these two problems by analyzing a GREEDY algorithm which approximately maximizes objective functions of the form Ψ(f, g), where f, g are two non-negative, monotone, submodular functions and Ψ is a quasiconvex 2-variables function, which is non decreasing with respect to the first variable. For the choice Ψ(f, g) , f/g, we recover RS, and for the choice Ψ(f, g) , f − g, we recover DS. To the best of our knowledge, this greedy approach is new for DS optimization. For RS optimization, it reduces to the standard GREEDRATIO algorithm that has already been analyzed in Bai et al. (2016). However, our analysis is novel for this case.

[1]  Gérard Cornuéjols,et al.  Submodular set functions, matroids and the greedy algorithm: Tight worst-case bounds and some generalizations of the Rado-Edmonds theorem , 1984, Discret. Appl. Math..

[2]  Rishabh K. Iyer,et al.  Curvature and Optimal Algorithms for Learning and Minimizing Submodular Functions , 2013, NIPS.

[3]  Michel Minoux,et al.  Accelerated greedy algorithms for maximizing submodular set functions , 1978 .

[4]  Rishabh K. Iyer,et al.  Algorithms for Optimizing the Ratio of Submodular Functions , 2016, ICML.

[5]  Yuval Filmus,et al.  A Tight Combinatorial Algorithm for Submodular Maximization Subject to a Matroid Constraint , 2012, 2012 IEEE 53rd Annual Symposium on Foundations of Computer Science.

[6]  Wei Chen,et al.  Scalable influence maximization for prevalent viral marketing in large-scale social networks , 2010, KDD.

[7]  Niv Buchbinder,et al.  Submodular Functions Maximization Problems , 2018, Handbook of Approximation Algorithms and Metaheuristics.

[8]  Yang Yu,et al.  Optimizing Ratio of Monotone Set Functions , 2017, IJCAI.

[9]  Jan Vondrák,et al.  Optimal approximation for submodular and supermodular optimization with bounded curvature , 2013, SODA.

[10]  E. Terzi,et al.  Team Formation: Striking a Balance between Coverage and Cost , 2020, 2002.07782.

[11]  J. Vondrák Submodularity and curvature : the optimal algorithm , 2008 .

[12]  Jitendra Malik,et al.  Normalized cuts and image segmentation , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[13]  Francis R. Bach,et al.  Learning with Submodular Functions: A Convex Optimization Perspective , 2011, Found. Trends Mach. Learn..

[14]  Grigorios Loukides,et al.  Limiting the Influence to Vulnerable Users in Social Networks: A Ratio Perspective , 2019, AINA.

[15]  Hui Lin,et al.  A Class of Submodular Functions for Document Summarization , 2011, ACL.

[16]  Vahab S. Mirrokni,et al.  Approximating submodular functions everywhere , 2009, SODA.

[17]  Samir Khuller,et al.  The Budgeted Maximum Coverage Problem , 1999, Inf. Process. Lett..

[18]  Dachuan Xu,et al.  Minimizing Ratio of Monotone Non-submodular Functions , 2019, Journal of the Operations Research Society of China.

[19]  Andreas Krause,et al.  Differentiable Submodular Maximization , 2018, IJCAI.

[20]  Washio Takashi,et al.  Prismatic algorithm for discrete D.C. programming problem , 2011 .

[21]  Moran Feldman,et al.  Guess Free Maximization of Submodular and Linear Sums , 2018, Algorithmica.

[22]  Siegfried Schaible,et al.  Fractional Programming , 2009, Encyclopedia of Optimization.

[23]  Jon Kleinberg,et al.  Maximizing the spread of influence through a social network , 2003, KDD '03.

[24]  Amin Karbasi,et al.  Submodular Maximization Beyond Non-negativity: Guarantees, Fast Algorithms, and Applications , 2019, ICML.

[25]  Jeff A. Bilmes,et al.  A Submodular-supermodular Procedure with Applications to Discriminative Structure Learning , 2005, UAI.

[26]  Zheng Wen,et al.  Online Influence Maximization under Independent Cascade Model with Semi-Bandit Feedback , 2016, NIPS.

[27]  Andreas Krause,et al.  Near-optimal Nonmyopic Value of Information in Graphical Models , 2005, UAI.

[28]  Éva Tardos,et al.  Maximizing the Spread of Influence through a Social Network , 2015, Theory Comput..

[29]  Rishabh K. Iyer,et al.  Algorithms for Approximate Minimization of the Difference Between Submodular Functions, with Applications , 2012, UAI.

[30]  I. Stancu-Minasian Nonlinear Fractional Programming , 1997 .

[31]  Vianney Perchet,et al.  Exploiting structure of uncertainty for efficient matroid semi-bandits , 2019, ICML.

[32]  Alexander Schrijver,et al.  Combinatorial optimization. Polyhedra and efficiency. , 2003 .

[33]  Satoru Iwata,et al.  A combinatorial strongly polynomial algorithm for minimizing submodular functions , 2001, JACM.

[34]  Vineeth N Balasubramanian,et al.  Submodular Batch Selection for Training Deep Neural Networks , 2019, IJCAI.