Abstract:
One of the basic primitives in the class of submodular optimization problems is
the submodular maximization under a cardinality constraint. Here we are given a ground set that is endowed with a monotone submodular function
and a parameter and the goal is to
return an optimal set of at most elements, i.e., is maximum among all subsets of of size at most . This basic primitive has many applications in machine learning as well as combinatorial optimization.
Example applications are agglomerative clustering, exemplar-based clustering,
categorical feature compression, document and corpus summarization,
recommender systems, search result diversification, data subset selection,
minimum spanning tree, max flow, global minimum cut, maximum
matching, traveling salesman problem, max clique, max cut, set cover and knapsack, among the others. In this paper, we propose the first dynamic algorithm for this problem. Given a stream of inserts and deletes of elements of an underlying ground set , we develop a dynamic algorithm that with high probability, maintains a -approximation of a cardinality-constrained monotone submodular maximization for any sequence of updates (inserts and deletes) in time , where is the maximum size of at any time. That is, the amortized update time of our algorithm is .
Chat is not available.