Space- and Time-Efficient Algorithm for Maintaining Dense Subgraphs on One-Pass Dynamic Streams

Abstract

While in many graph mining applications it is crucial to handle a stream of updates efficiently in terms of both time and space, not much was known about achieving such type of algorithm. In this paper we study this issue for a problem which lies at the core of many graph mining applications called densest subgraph problem. We develop an algorithm that achieves time- and space-efficiency for this problem simultaneously. It is one of the first of its kind for graph problems to the best of our knowledge. Given an input graph, the densest subgraph is the subgraph that maximizes the ratio between the number of edges and the number of nodes. For any ε&gt;0, our algorithm can, with high probability, maintain a (4+ε)-approximate solution under edge insertions and deletions using ~O(n) space and ~O(1) amortized time per update; here, $n$ is the number of nodes in the graph and ~O hides the O(polylog_{1+ε} n) term. The approximation ratio can be improved to (2+ε) with more time. It can be extended to a (2+ε)-approximation sublinear-time algorithm and a distributed-streaming algorithm. Our algorithm is the first streaming algorithm that can maintain the densest subgraph in one pass. Prior to this, no algorithm could do so even in the special case of an incremental stream and even when there is no time restriction. The previously best algorithm in this setting required O(log n) passes [BahmaniKV12]. The space required by our algorithm is tight up to a polylogarithmic factor.<p>QC 20150811</p

Unifying and Strengthening Hardness for Dynamic Problems via the Online Matrix-Vector Multiplication Conjecture

“…(1) Generally speaking, for most previous hardness results in [36,1,29] that rely on various conjectures, except those relying on the Strong Exponential Time Hypothesis (SETH), our OMv conjecture implies hardness bounds on the amortized time per operation that are the same or better. (2) We also obtain new results such as those for vertex color distance oracles (studied in [26,11] and used to tackle the minimum Steiner tree problem [31]), restricted top trees with edge query problem (used to tackle the minimum cut problem in [16]), and the dynamic densest subgraph problem [7]. (3) Some minor improvement can in fact immediately be obtained since our conjecture implies a very strong bound for Pǎtraşcu's multiphase problem [36], giving improved bounds for many problems considered in [36].…”

The K-clique Densest Subgraph Problem

“…When restrictions on the size of S are imposed the problem becomes NP-hard [40]. Finally, the densest subgraph problem has been considered in various settings, including MapReduce [10], the streaming [10], the dynamic setting [21,45] and their combination recently [13]. Triangle Counting and Listing.…”

Clustered Hierarchical Entropy-Scaling Search of Astronomical and Biological Data

“…Given the rise of "big data," there has been an interest in sublinear-time algorithms [17], [18], [19]. With hierarchical entropy-scaling search, we demonstrate not just a sublineartime search algorithm (one whose advantages over naïve search will grow as data sets grow) but a search algorithm whose asymptotic complexity is not a function of data set size but rather geometric properties of the data.…”