Intrinsic Degree: An Estimator of the Local Growth Rate in Graphs

The neighborhood size of a query node in a graph often grows exponentially with the distance to the node, making a neighborhood search prohibitively expensive even for small distances. Estimating the growth rate of the neighborhood size is therefore an important task in order to determine an appropriate distance for which the number of traversed nodes during the search will be feasible. In this work, we present the intrinsic degree model, which captures the growth rate of exponential functions through the analysis of the infinitesimal vicinity of the origin. We further derive an estimator which allows to apply the intrinsic degree model to graphs. In particular, we can locally estimate the growth rate of the neighborhood size by observing the close neighborhood of some query points in a graph. We evaluate the performance of the estimator through experiments on both artificial and real networks.

[1]  James Bailey,et al.  Measuring dependency via intrinsic dimensionality , 2016, 2016 23rd International Conference on Pattern Recognition (ICPR).

[2]  Ken-ichi Kawarabayashi,et al.  Estimating Local Intrinsic Dimensionality , 2015, KDD.

[3]  Michael E. Houle,et al.  Local Intrinsic Dimensionality II: Multivariate Analysis and Distributional Support , 2017, SISAP.

[4]  James Bailey,et al.  The vulnerability of learning to adversarial perturbation increases with intrinsic dimensionality , 2017, 2017 IEEE Workshop on Information Forensics and Security (WIFS).

[5]  Michael E. Houle,et al.  Effective and Efficient Algorithms for Flexible Aggregate Similarity Search in High Dimensional Spaces , 2015, IEEE Transactions on Knowledge and Data Engineering.

[6]  Michael E. Houle,et al.  Efficient Algorithms for Similarity Search in Axis-Aligned Subspaces , 2014, SISAP.

[7]  D. Bu,et al.  Topological structure analysis of the protein-protein interaction network in budding yeast. , 2003, Nucleic acids research.

[8]  Derek de Solla Price,et al.  A general theory of bibliometric and other cumulative advantage processes , 1976, J. Am. Soc. Inf. Sci..

[9]  Michael E. Houle,et al.  Query Expansion for Content-Based Similarity Search Using Local and Global Features , 2017, ACM Trans. Multim. Comput. Commun. Appl..

[10]  Michael E. Houle,et al.  Dimensional Testing for Multi-step Similarity Search , 2012, 2012 IEEE 12th International Conference on Data Mining.

[11]  Béla Bollobás,et al.  Random Graphs , 1985 .

[12]  James Bailey,et al.  Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality , 2018, ICLR.

[13]  Emily B. Fox,et al.  Sparse graphs using exchangeable random measures , 2014, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[14]  Yiming Yang,et al.  The Enron Corpus: A New Dataset for Email Classi(cid:12)cation Research , 2004 .

[15]  Hisashi Kashima,et al.  Generalized Expansion Dimension , 2012, 2012 IEEE 12th International Conference on Data Mining Workshops.

[16]  Albert,et al.  Emergence of scaling in random networks , 1999, Science.

[17]  Peer Kröger,et al.  Dimensional Testing for Reverse k-Nearest Neighbor Search , 2017, Proc. VLDB Endow..

[18]  James Bailey,et al.  Dimensionality-Driven Learning with Noisy Labels , 2018, ICML.

[19]  Michael E. Houle,et al.  Improving k-NN Graph Accuracy Using Local Intrinsic Dimensionality , 2017, SISAP.

[20]  Albert,et al.  Topology of evolving networks: local events and universality , 2000, Physical review letters.

[21]  Le Song,et al.  Iterative Learning with Open-set Noisy Labels , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[22]  Michael E. Houle,et al.  Dimensionality, Discriminability, Density and Distance Distributions , 2013, 2013 IEEE 13th International Conference on Data Mining Workshops.

[23]  Michael E. Houle,et al.  Local Intrinsic Dimensionality I: An Extreme-Value-Theoretic Foundation for Similarity Applications , 2017, SISAP.

[24]  Stanley Milgram,et al.  An Experimental Study of the Small World Problem , 1969 .