Network entropy explained

In network science, the network entropy is a disorder measure derived from information theory to describe the level of randomness and the amount of information encoded in a graph.^[1] It is a relevant metric to quantitatively characterize real complex networks and can also be used to quantify network complexity^[2]

Formulations

According to a 2018 publication by Zenil et al. there are several formulations by which to calculate network entropy and, as a rule, they all require a particular property of the graph to be focused, such as the adjacency matrix, degree sequence, degree distribution or number of bifurcations, what might lead to values of entropy that aren't invariant to the chosen network description.

Degree Distribution Shannon Entropy

The Shannon entropy can be measured for the network degree probability distribution as an average measurement of theheterogeneity of the network.

l{H} = - \sum ^{N -1} _{k =1} P(k) ln {P(k)}

This formulation has limited use with regards to complexity, information content, causation and temporal information. Be that as it may, algorithmic complexity has the ability to characterize any general or universal property of a graph or network and it is proven that graphs with low entropy have low algorithmic complexity because the statistical regularities found in a graph are useful for computer programs to recreate it. The same cannot be said for high entropy networks though, as these might have any value for algorithmic complexity.^[3]

Random Walker Shannon Entropy

Due to the limits of the previous formulation, it is possible to take a different approach while keeping the usage of the original Shannon Entropy equation.

Consider a random walker that travels around the graph, going from a node

i to any node j adjacent to i with equal probability. The probability distribution p _{ij} that describes the behavior of this random walker would thus bep _{ij} = \begin{cases} (1)/(k _{i}) , & if A _{ij} = 1 \\ 0 , & if A _{ij} = 0 \\ \end{cases} ,

where

(A _{ij} ) is the graph adjacency matrix and k _{i} is the node i degree.

From that, the Shannon entropy from each node

l{S} _{i} can be defined asl{S} _{i} = - \sum ^{N -1} _{j =1} p _{ij} ln {p _{ij} } = \ln

and, since

max(k _{i)} = N - 1 , the normalized node entropy l{H} _{i} is calculatedl{H} _{i} = (l{S)/(_{i}{max(l{S}}) _{i)}} = (ln {k _{i})/() } = \frac

This leads to a normalized network entropy

l{H} , calculated by averaging the normalized node entropy over the whole network:^[4] l{H} = (1)/(N) \sum ^{N} _{i =1} l{H} _{i} = (1)/(N ln (N -1 )) \sum ^{N} _{i =1} ln {k _{i}}

The normalized network entropy is maximal

l{H} = 1 when the network is fully connected and decreases the sparser the network becomes l{H} = 0 . Notice that isolated nodes k _{i} = 0 do not have its probability p _{ij} defined and, therefore, are not considered when measuring the network entropy. This formulation of network entropy has low sensitivity to hubs due to the logarithmic factor and is more meaningful for weighted networks.,^[4] what ultimately makes it hard to differentiate scale-free networks using this measure alone.^[2]

Random Walker Kolmogorov–Sinai Entropy

The limitations of the random walker Shannon entropy can be overcome by adapting it to use a Kolmogorov–Sinai entropy. In this context, network entropy is the entropy of a stochastic matrix associated with the graph adjacency matrix

(A _{ij} ) and the random walker Shannon entropy is called the dynamic entropy of the network. From that, let λ be the dominant eigenvalue of (A _{ij} ) . It is proven that ln λ satisfies a variational principal^[5] that is equivalent to the dynamic entropy for unweighted networks, i.e., the adjacency matrix consists exclusively of boolean values. Therefore, the topological entropy is defined asl{H} = ln λ

This formulation is important to the study of network robustness, i.e., the capacity of the network to withstand random structuralchanges. Robustness is actually difficult to be measured numerically whereas the entropy can be easily calculated for any network, which is especially important in the context of non-stationary networks. The entropic fluctuation theorem shows that this entropy is positively correlated to robustness and hence a greater insensitivity of an observable to dynamic or structural perturbations of the network. Moreover, the eigenvalues are inherently related to the multiplicity of internal pathways, leading to a negative correlation between the topological entropy and the shortest average path length.^[6]

Other than that, the Kolmogorov entropy is related to the Ricci curvature of the network,^[7] a metric that has been used to differentiate stages of cancer from gene co-expression networks,^[8] as well as to give hallmarks of financial crashes from stock correlation networks^[9]

Von Neumann entropy

\rho : historically, the first proposed candidate for such a density matrix has been an expression of the Laplacian matrix L associated with the network. The average von Neumann entropy of an ensemble is calculated as:^[10]

$$\_\; =\; -\backslash langle\backslash mathrm\backslash rho\backslash log(\backslash rho)\backslash rangle$$

For random network ensemble

G(N,p) , the relation between S _{VN} and S is nonmonotonic when the average connectivity p(N -1 ) is varied.

For canonical power-law network ensembles, the two entropies are linearly related.^[11]

$$\_\; =\; \backslash eta\; +\; \backslash beta$$

Networks with given expected degree sequences suggest that, heterogeneity in the expected degree distribution implies an equivalence between a quantum and a classical description of networks, which respectively corresponds to the von Neumann and the Shannon entropy.^[12]

This definition of the Von Neumann entropy can also be extended to multilayer networks with tensorial approach^[13] and has been used successfully to reduce their dimensionality from a structural point of perspective.^[14]

However, it has been shown that this definition of entropy does not satisfy the property of sub-additivity (see Von Neumann entropy's subadditivity), expected to hold theoretically. A more grounded definition, satisfying this fundamental property, has been introduced by Manlio De Domenico and Biamonte^[15] as a quantum-like Gibbs state

$$\backslash rho(\backslash beta)=\backslash frac$$

where $$Z(\backslash beta)=Tr[e^\{-\backslash beta\; L\}]$$ is a normalizing factor which plays the role of the partition function, and

\beta is a tunable parameter which allows multi-resolution analysis. If \beta is interpreted as a temporal parameter, this density matrix is formally proportional to the propagator of a diffusive process on the top of the network.

This feature has been used to build a statistical field theory of complex information dynamics, where the density matrix can be interpreted in terms of the super-position of streams operators whose action is to activate information flows among nodes.^[16] The framework has been successfully applied to analyze the protein-protein interaction networks of virus-human interactomes, including the SARS-CoV-2, to unravel the systemic features of infection of the latter at microscopic, mesoscopic and macroscopic scales,^[17] as well as to assess the importance of nodes for integrating information flows within the network and the role they play in network robustness.^[18]

This approach has been generalized to deal with other types of dynamics, such as random walks, on the top of multilayer networks, providing an effective way to reduce the dimensionality of such systems without altering their structure.^[19] Using both classical and maximum-entropy random walks, the corresponding density matrices have been used to encode the network states of the human brain and to assess, at multiple scales, connectome’s information capacity at different stages of dementia.^[20]

Maximum Entropy Principle

See main article: Principle of maximum entropy. The maximum entropy principle is a variational principal stating that the probability distribution best representing the current state of a system is the one which maximizes the Shannon entropy.^[21] This concept can be used to generate an ensemble of random graphs with given structural properties derived from the maximum entropy approach which, in its turn, describes the most probable network configuration: the maximum entropy principle allows for maximally unbiased information when lacking complete knowledge (microscopic configuration is not accessible, e.g.: we don't know the adjacency matrix). On the other hand, this ensemble serves as a null model when the actual microscopic configuration of the network is known, allowing to assess the significance of empirical patterns found in the network^[22]

Network Ensembles

See main article: Entropy of network ensembles. It is possible to extend the network entropy formulations to instead measure the ensemble entropy. Let

\Omega be an ensemble of all graphs G with the same number N of nodes and type of links, P(G) is defined as the occurrence probability of a graph G \in \Omega . From the maximum entropy principle, it is desired to achieve P(G) such that it maximizes the Shannon entropyl{H} = - \sum _{G} P(G) ln P(G)

Moreover, constraints shall be imposed in order to extract conclusions. Soft constraints (constraints set over the whole ensemble

\Omega ) will lead to the canonical ensemble whereas hard constraints (constraints set over each graph G ) are going to define the microcanonical ensemble. In the end, these ensembles lead to models that can validate a range of network patterns and even reconstruct graphs from limited knowledge, showing that maximum entropy models based on local constraints can quantify the relevance of a set of observed features and extracting meaningful information from huge big data streams in real-time and high-dimensional noisy data.

Notes and References

1. Anand . Kartik . Krioukov . Dmitri . Bianconi . Ginestra . Entropy distribution and condensation in random networks with a given degree distribution . Physical Review E . 2014 . 89 . 6 . 062807 . 10.1103/PhysRevE.89.062807 . 25019833 . 1403.5884 . 2014PhRvE..89f2807A . 761765 . anand2014entropy.
2. freitas2019detailed. Freitas . Cristopher GS . Aquino . Andre LL . Ramos . Heitor S . Frery . Alejandro C . Rosso . Osvaldo A . A detailed characterization of complex networks using Information Theory . Scientific Reports . 2019 . 9 . 1 . 16689 . 10.1038/s41598-019-53167-5 . 31723172 . 6853913 . 2019NatSR...916689F . 207987035 .
3. Zenil . Hector . Kiani . Narsis A . Tegnér . Jesper . A review of graph and network complexity from an algorithmic information perspective . Entropy . 2018 . 20 . 8 . 551 . 10.3390/e20080551 . 33265640 . 7513075 . 2018Entrp..20..551Z . zenil2018review. free .
4. Book: Small . Michael . 2013 IEEE International Symposium on Circuits and Systems (ISCAS2013) . Complex networks from time series: Capturing dynamics . 2013 . 2509–2512 . 10.1109/ISCAS.2013.6572389 . 978-1-4673-5762-3 . 9275909 . https://ieeexplore.ieee.org/document/6572389 . small2013complex.
5. Arnold . Ludwig . Gundlach . Volker Matthias . Demetrius . Lloyd . Evolutionary formalism for products of positive random matrices . The Annals of Applied Probability . 1994 . 4 . 3 . 859–901. 10.1214/aoap/1177004975 . 2245067 . free .
6. Demetrius . Lloyd . Manke . Thomas . Robustness and network evolution—an entropic principle . Physica A: Statistical Mechanics and Its Applications . 2005 . 346 . 3 . 682–696. 10.1016/j.physa.2004.07.011 . 2005PhyA..346..682D .
7. Lott . J. . Villani . C. . Ricci curvature for metric-measure spaces via optimal transport . Annals of Mathematics . 2009. 169 . 3 . 903–991 . 10.4007/annals.2009.169.903 . 15556613 . math/0412127 .
8. Sandhu . R. . Georgiou . T. . Reznik . E. . Zhu . L. . Kolesov . I. . Senbabaoglu . Y. . Tannenbaum . A. . Graph curvature for differentiating cancer networks . Scientific Reports . 2015. 5 . 12323 . 10.1038/srep12323 . 26169480 . 4500997 . 2015NatSR...512323S .
9. Sandhu . Romeil S . Georgiou . Tryphon T . Tannenbaum . Allen R . Ricci curvature: An economic indicator for market fragility and systemic risk . Science Advances . 2016. 2 . 5 . e1501495 . 10.1126/sciadv.1501495 . 27386522 . 4928924 . 2016SciA....2E1495S .
10. Du . Wenxue . Li . Xueliang . Li . Yiyang . Severini . Simone . A note on the von Neumann entropy of random graphs . Linear Algebra and Its Applications . 30 December 2010 . 433 . 11 . 1722–1725 . 10.1016/j.laa.2010.06.040 . en . 0024-3795. free .
11. Anand . Kartik . Bianconi . Ginestra . Entropy measures for networks: Toward an information theory of complex topologies . Physical Review E . 13 October 2009 . 80 . 4 . 045102 . 10.1103/PhysRevE.80.045102 . 19905379 . 0907.1514 . 2009PhRvE..80d5102A . 27419558 .
12. Anand . Kartik . Bianconi . Ginestra . Severini . Simone . Shannon and von Neumann entropy of random networks with heterogeneous expected degree . Physical Review E . 18 March 2011 . 83 . 3 . 036109 . 10.1103/PhysRevE.83.036109 . 21517560 . 1011.1565 . 2011PhRvE..83c6109A . 1482301 .
13. De Domenico . Manlio . Solé-Ribalta . Albert . Cozzo . Emanuele . Kivelä . Mikko . Moreno . Yamir . Porter . Mason A. . Gómez . Sergio . Arenas . Alex . Mathematical Formulation of Multilayer Networks . Physical Review X . 4 December 2013 . 3 . 4 . 041022 . 10.1103/PhysRevX.3.041022 . 1307.4977 . 2013PhRvX...3d1022D . 16611157 .
14. De Domenico . Manlio . Nicosia . Vincenzo . Arenas . Alex . Latora . Vito . Structural reducibility of multilayer networks . Nature Communications . 23 April 2015 . 6 . 6864 . 10.1038/ncomms7864 . 25904309 . 2015NatCo...6.6864D . 16776349 .
15. De Domenico . Manlio . Biamonte . Jacob. Spectral Entropies as Information-Theoretic Tools for Complex Network Comparison . Physical Review X . 21 December 2016 . 6 . 4 . 041062 . 10.1103/PhysRevX.6.041062. 1609.01214 . 2016PhRvX...6d1062D . 51786781 .
16. Ghavasieh . Arsham. Nicolini . Carlo. De Domenico . Manlio. Statistical physics of complex information dynamics . Physical Review E . 10 November 2020 . 102 . 5 . 052304 . 10.1103/PhysRevE.102.052304. 33327131 . 2010.04014 . 2020PhRvE.102e2304G . 222208856 .
17. Ghavasieh . Arsham. Bontorin . Sebastiano. Artime . Oriol. Verstraete . Nina. De Domenico . Manlio. Multiscale statistical physics of the pan-viral interactome unravels the systemic nature of SARS-CoV-2 infections . Communications Physics . 23 April 2021 . 4 . 1 . 83 . 10.1038/s42005-021-00582-8. 2008.09649. 2021CmPhy...4...83G . free .
18. Ghavasieh . Arsham. Stella . Massimo. Biamonte . Jacob. De Domenico . Manlio. Unraveling the effects of multiscale network entanglement on empirical systems. Communications Physics . 10 June 2021 . 4 . 1 . 129 . 10.1038/s42005-021-00633-0. 2008.05368. 2021CmPhy...4..129G . 221104066 .
19. Ghavasieh . Arsham. De Domenico . Manlio. Enhancing transport properties in interconnected systems without altering their structure . Physical Review Research . 13 February 2020 . 2 . 1 . 13–15 . 10.1103/PhysRevResearch.2.013155. 2001.04450 . 2020PhRvR...2a3155G . 210165034 .
20. Benigni . Barbara. Ghavasieh . Arsham. Corso . Alessandra. D'Andrea . Valeria. De Domenico . Manlio. Persistence of information flow: a multiscale characterization of human brain . Network Neuroscience . 22 June 2021 . 5. 3 . 831–850 . 10.1162/netn_a_00203. 34746629 . 8567833 . free .
21. Jaynes . E. T. . Edwin Thompson Jaynes . 1957 . Information Theory and Statistical Mechanics . Physical Review . Series II . 106 . 4 . 620–630 . 10.1103/PhysRev.106.620 . 87305. 1957PhRv..106..620J . 17870175 .
22. Cimini . Giulio . Squartini . Tiziano . Saracco . Fabio . Garlaschelli . Diego . Gabrielli . Andrea . Caldarelli . Guido . The statistical physics of real-world networks . Nature Reviews Physics . 2019 . 1 . 1 . 58–71. 10.1038/s42254-018-0002-6 . 1810.05095 . 2019NatRP...1...58C . 52963395 .