GUISE: a uniform sampler for constructing frequency histogram of graphlets

In large networks, the connected triples are useful for solving various tasks including link prediction, community detection, and spam filtering. Existing works in this direction concern mostly with the exact or approximate counting of connected triples that are closed (aka, triangles). Evidently, the task of triple sampling has not been explored in depth, although sampling is a more fundamental task than counting, and the former is useful for solving various other tasks, including counting. In recent years, some works on triple sampling have been proposed that are based on direct sampling, solely for the purpose of triangle count approximation. They sample only from a uniform distribution, and are not effective for sampling triples from an arbitrary user-defined distribution. In this work we present two indirect triple sampling methods that are based on Markov Chain Monte Carlo (MCMC) sampling strategy. Both of the above methods are highly efficient (several magnitudes faster on large datasets) compared to a direct sampling-based method, specifically for the task of sampling from a non-uniform probability distribution. Another significant advantage of the proposed methods is that they can sample triples from networks that have restricted access, on which a direct sampling based method is simply not applicable.

show abstract

“…For example, sampling has been used for finding interesting subgraphs [1], communities [11], and graphlet frequency distribution [15].…”

Section: Related Workmentioning

confidence: 99%

“…Bhuiyan et al [15] use it for sampling graphlets, Maiya et al [11] use it for sampling community structure, and Gjoka et al [8] use it for finding an approximate degree distribution of Facebook network. However, each of these works have a different objective and they sample from different population.…”

Section: Related Workmentioning

confidence: 99%

Sampling Triples from Restricted Networks using MCMC Strategy

Rahman

Hasan

2014

Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management

Self Cite

View full text Add to dashboard Cite

show abstract

“…Formally, given a graph G, graphlets of G are small, connected, non-isomorphic, and vertex-induced subgraphs of G. Although earlier works [20,21,22] on graphlets focus on undirected graphs, we consider directed graphlets to preserve the inherent directionality of FCGs. Figure 1 enumerates all the 13 unique types of (directed) graphlets ω 3,i 2 (i = 1, 2, .…”

Section: B Graphletsmentioning

confidence: 99%

“…However, the real problem is that we need to uniformly sample graphlets from the FCG without enumerating the sample space. Fortunately, two recent advances on graph mining, GRAFT [21] and GUISE [22], show that GFD can be estimated without enumerating all graphlets. Inspired by these works, we use MCMC to sample the directed FCG.…”

Section: B Efficient Gfd Estimationmentioning

confidence: 99%

See 1 more Smart Citation

ACTS: Extracting Android App topologiCal signature through graphleT Sampling

Peng

Gao

Sisodia

et al. 2016

2016 IEEE Conference on Communications and Network Security (CNS)

View full text Add to dashboard Cite

Abstract-Android systems are widely used in mobile & wireless distributed systems. In the near future, Android is believed to dominate the mobile distributed environment. However, with the popularity of Android-based smartphones/tablets comes the rampancy of Android-based malware. In this paper, we propose a novel topological signature of Android apps based on the function call graphs (FCGs) extracted from their Android App PacKages (APKs). Specifically, by leveraging recent advances in graphlet sampling, the proposed method fully captures the invocatorinvocatee relationship at local neighborhoods in an FCG without exponentially inflating the state space. Using real benign app and malware samples, we demonstrate that our method, ACTS (App topologiCal signature through graphleT Sampling), can detect malware and identify malware families robustly and efficiently.More importantly, we demonstrate that, without augmenting the FCG with any semantic features such as bytecode-based vertex typing, local topological information captured by ACTS alone can achieve a high malware detection accuracy. Since ACTS only uses structural features, which are orthogonal to semantic features, it is expected that combining them would give a greater improvement in malware detection accuracy than combining nonorthogonal semantic features.

show abstract