Access to this full-text is provided by Springer Nature.
Content available from Journal of Cloud Computing
This content is subject to copyright. Terms and conditions apply.
Journal of Cloud Computing:
Advances, Systems and Applications
Fan et al. Journal of Cloud Computing: Advances, Systems
and Applications (2020) 9:24
https://doi.org/10.1186/s13677-020-00169-8
RESEARCH Open Access
Revisiting the power of reinsertion for
optimal targets of network attack
Changjun Fan1*†, Li Zeng1†, Yanghe Feng1,BaoxinXiu
2, Jincai Huang1and Zhong Liu1
Abstract
Understanding and improving the robustness of networks has significant applications in various areas, such as
bioinformatics, transportation, critical infrastructures, and social networks. Recently, there has been a large amount of
work on network dismantling, which focuses on removing an optimal set of nodes to break the network into small
components with sub-extensive sizes. However, in our experiments, we found these state-of-the-art methods,
although seemingly different, utilize the same refinement technique, namely reinsertion, to improve the performance.
Despite being mentioned with understatement, the technique essentially plays the key role in the final performance.
Without reinsertion, the current best method would deteriorate worse than the simplest heuristic ones; while with
reinsertion, even the random removal strategy achieves on par with the best results. As a consequence, we, for the
first time, systematically revisit the power of reinsertion in network dismantling problems. We re-implemented and
compared 10 heuristic and approximate competing methods on both synthetic networks generated by four classical
network models, and 18 real-world networks which cover seven different domains with varying scales. The
comprehensive ablation results show that: i) HBA (High Betweenness Adaption, no reinsertion) is the most effective
network dismantling strategy, however, it can only be applicable in small scale networks; ii) HDA (High Degree
Adaption, with reinsertion) achieves the best balance between effectiveness and efficiency; iii) The reinsertion
techniques help improve the performance for most current methods; iv) The one, which adds back the node based
on that it joins the clusters minimizing the multiply of both numbers and sizes, is the most effective reinsertion
strategy for most methods. Our results can be a survey reference to help further understand the current methods and
thereafter design the better ones.
Keywords: Network dismantling, Reinsertion
Introduction
Many real-world systems can be described through the
complex network perspective, including air transport [17],
power grid [3], malicious organization [9,10], Internet [3]
or inter-personal networks [15]. One of the most impor-
tant topics on these networks is about the robustness, i.e.,
the capacity to maintain the functionality after a major
failure [29]. Since connectivity is the fundamental basic
for almost all behaviors on networks, researches thus try
to quantify how the connectivity is affected by node(or
*Correspondence: fanchangjun09@163.com
†Changjun Fan and Li Zeng contributed equally to this work.
1College of Systems Engineering, National University of Defense Technology,
Changsha, Hunan, China
Full list of author information is available at the end of the article
link) removal, and there comes with the well-defined net-
work dismantling problem [1], which aims at identifying
an optimal sequence of nodes that maximizes the dam-
age on the network connectivity [5]. Such analysis yields
a wide range of practical applications, such as immu-
nize the epidemic propagation in populations [23], block
the rumor spreading on social networks [15], prevent the
virus diffusion in computer networks [7], etc.
However, the exact solution is computationally
intractable for medium and large networks due to its
NP-hard nature [5], thus a large number of approximate
methods have been proposed, including the heuristic
methods [4,11,12,21,23,31], and some message-passing
algorithms [5,22]. The former methods often greedily
© The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit
to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The
images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 2 of 13
select target nodes based on local metrics, like node
degree, which often leads to sub-optimal solutions; the
latter ones are more accurate and global, while they need
to iterate certain steps on the whole network to select the
suitable candidate nodes [31], which would sacrifice some
efficiency.
Although these methods looks different from each
other, many of them [5,21,22,24,31]sharethesame
refinement technique, named reinsertion (we later intro-
duce it in detail in Section 2), which is just simply
mentioned in the respective literature, while has sig-
nificant influence on the final results. As illustrated in
Fig. 1, we draw the robustness curves (“Robustness mea-
sure” section) of random removal, simplest heuristic HDA
and the representative CI (details of these methods will
be introduced in “Competing methods” section) on a real-
world Gnutella31 network [18]. We can see that without
reinsertion, the representative method CI cannot even
beat the simplest heuristic HDA, while with reinsertion,
the random removal strategy can achieve comparable
performance than the state-of-the-art results. In some
literature, people just compare their methods enhanced
with reinsertion with others without reinsertion, and then
report the ’fake’ superiority of their model, since we are
not sure whether the superiority comes from the model
itself or just the reinsertion. Such confused results pre-
vent us from selecting the best algorithm to handle the
application at hand.
In this paper, we systematically investigate the power of
reinsertion on the current methods for network disman-
tling. As far as we know, there are no previous efforts
that conduct such comprehensive ablation studies for the
reinsertion. We aim at figuring out the following three
questions: i) Which is the current best method if all without
reinsertion? ii)Which one is the best if all with reinsertion?
iii) Which is the best reinsertion strategy?
To achieve this, we conduct ablation study
(with/without reinsertion) for all the current network dis-
mantling methods, including both traditional heuristics
and the state-of-art message-passing ones, on synthetic
networks and real-world networks. We use four random
network models, including ER [8], WS [30], BA [3]and
PLC [13], to generate diverse graphs with varying sizes
and structures by controlling the model parameters. For
real-world networks, we select 18 real networks covering
7 domains and with different scales. Considering that the
network robustness can be described by different mea-
sures,wechoosetheareaundertherobustnesscurveas
themainevaluationmetric,sinceitcapturestheresponse
of the whole dismantling process. Extensive experi-
ments demonstrate that the reinsertion can significantly
improve the performance regardless of the network types
and the methods. Besides, since reinsertion is rather
effective for the network dismantling problem, perhaps
people should focus on this technique itself rather than
otheraspects,soastodesignabetterattackstrategy.
The main contributions of this paper are summarized as
follows:
1. We conduct comprehensive ablation studies that are
with and without reinsertion for the network
dismantling problem. We compare 10 competing
Fig. 1 Illustration of the reinsertion power on Gnutella31 network
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 3 of 13
methods on both synthetic graphs generated from
four random network types and 18 real-world
networks covering seven domains and scales up to
hundreds of thousands nodes;
2. We design two other reinsertion strategies, and
empirically prove that they have surpassed the
previous reinsertion technique in a large margin;
3. The results obtained in this paper could provide a
valuable guide for selecting and designing the most
appropriate method for practical network
dismantling problems.
The rest of the paper is organized as follows. We
introduce the reinsertion method, robustness mea-
sures, competing methods and experimental data in
“Method” section. We analyze the comprehensive abla-
tion results and effects of different reinsertion strategies
in “Results” section. Finally, we conclude the paper in
Section 5.
Method
In this section, we introduce the experimental setups.
We first introduce the robustness measure to evaluate
the dismantling efficacy, then we introduce the reinser-
tion technique that is widely adopted in most current
competitors. After that, we describe the competitors we
are to analyze and the experimental data, including both
synthetic graphs and real-world networks.
Robustness measure
Network dismantling is to identify a sequence of nodes
of which removal would degrade the network connec-
tivity maximally, and this connectivity disintegration is
often measured as the relative reduction in the size of
the giant(largest) connected component (GCC size) [5,
21]. The smaller the remaining GCC size, the more the
network is considered to have been disintegrated.
We consider the area under the robustness curve as the
evaluation metric, which is plotted with horizontal axis
being the fraction of nodes removed, and the vertical axis
being the remaining GCC size. It is defined as:
R=1
N
N
Q=1
s(Q)(1)
where Nis the number of graph nodes, s(Q)is the remain-
ing GCC size after removing Qnodes. Intuitively this mea-
sure is equivalent to assessing how many nodes the GCC
contains when a new node is deleted from the network,
and sum this for all nodes [29]. Note that Eq. 1captures
the network’s response to the dismantling throughout the
whole process, and the computation of Rrequires a rank-
ing of the nodes, we are interested in minimizing Rover
all possible node orders.
In this paper, we evaluate the ablation performance of
reinsertion for this robustness measure.
Reinsertion technique
The reinsertion is firstly proposed as an independent
strategy for network destruction and immunization [27],
and later developed as an important refinement tech-
nique for other dismantling strategies. The reinsertion
starts from the point, where the network has been dis-
mantled over by a certain strategy, it adds back one of the
removed node, chosen such that, once reinserted, it joins
the smallest number of clusters. When the node is rein-
serted, restore the edges with its neighbors which are in
the network (but not the ones with neighbors not yet rein-
serted, if any). Repeat the above the procedure until all the
nodes are back in the network.
As is shown in Fig. 2, each node is assigned an index
c(i)given by the number of clusters it would join if it is
reinserted in the network. The red node has c(red)=2,
while the blue one has c(blue)=4, the green node has
c(green)=3. Then the node with the smallest c(i)is
reinserted, i.e., the red node. After that, the c(i)s are recal-
culated and the new node with smallest c(i)is found and
reinserted. Repeat these steps until the terminal criteria
meets.
We will later show with extensive experiments how
powerful such a simple technique is to the current net-
work dismantling methods.
Fig. 2 Illustration of the reinsertion procedure
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 4 of 13
Competing methods
In this paper, we compare with 9 most representative com-
peting methods. The first five are traditional heuristics
which are based on some local or global structure cen-
trality, such as degree, betweenness, closeness, pagerank,
or collective influence. The remaining five are specifically
designed for dismantling networks. Note that we also add
aRandom removal strategy as a worst possible baseline.
High Degree Adaptive (HDA) [23]. HDA is an adap-
tive version of high degree method [2]. Within each step,
the node with the highest degree is removed, and then the
remaining degrees are updated.
High Betweenness Adaptive (HBA) [12].HBAisthe
adaptive version of the high betweenness method, where
the betweenness centrality of the remaining nodes is
recomputed after each node removal. Betweenness cen-
trality of a node equals to the sum of the fraction of all
pairs shortest paths that pass through this node. It is a very
useful centrality measure that benefits many network-
related applications such as community detection and
network vulnerability. However, the high computing cost
prohibits its use in large-scale problem settings.
High Closeness Adaptive(HCA) [4]. HCA is the adap-
tive version of the high closeness method. Closeness cen-
trality describes how close a node is to all the other nodes
in the graph. It is calculated as the reciprocal of average
distances from one node to all the others. Similar as HBA,
the high complexity cost prevents its application in large
networks.
High PageRank Adaptive (HPRA) [6].HPRAisthe
adaptive version of high PageRank method. PageRank has
been widely employed in search engines, as it provides a
global ranking of all web pages, regardless of their content,
based solely on their location in the Web’s graph structure
[6]. PageRank computes the probabilities for a random-
walking agent to reach every node in the network, which
is also regarded as useful indications to supervise the
network attack.
Collective Influence(CI) [21]. The Collective Influence
measureisdefinedastheproductofthenode’sreduced
degree (i.e. original degree minus one) with the sum of the
reduced degrees of the nodes that are within a constant
hops away from it. This measure describes the propor-
tion of other nodes that can be reached from a given
node, assuming the nodes with higher CI values play
more crucial roles in networks. The CI method sequen-
tially removes the node with the highest CI value and
recalculating the collective influence for the rest following
operations.
MinSum [5]. MinSum is proposed to address the net-
work dismantling problem. It consists three stages, which
firstly utilizes a variant of message-passing algorithm to
break all the cycles, and then breaks the remaining tree
into small components by removing a fraction of nodes
that vanishes in the large size limit. In the third stage, it
greedily reinserts some nodes that close cycles without
increasing too much the size the largest component, to
reduce the total number of nodes removed.
Belief Propagation-guided Decimation (BPD) [22].
BPD is very similar as MinSum, which contains the same
three stages. The difference lies on that BPD treats the
decycling problem as the minimum-FVS construction.
The FVS refers to the feedback vertex set, which is a
set of node that will cause the network to become a for-
est if being deleted. To solve this problem, BPD proposes
a belief propagation-guided decimation algorithm. After,
it conducts the same subsequent steps, including tree
breaking and node reinsertion.
CoreHD [31]. CoreHD also contains the similar three
stages. The only difference lies in the decycling stage.
Unlike the message-passing or belief-propagation algo-
rithm, CoreHD instead seeks to remove the minimum
nodes to empty the 2-core subgraph in the network, since
the network is acyclic equals to that the 2-core subgraph is
empty. CoreHD greedily remove the highest degree node
inthe2-coresubgraphuntiltheend.
GND [25]. GND is the state-of-the-art method to
address the network dismantling problem with non-unit
removal costs. It first defines a node weighted Laplacian,
and then proposes a simple and elegant approximate algo-
rithm to calculate its second smallest eigenvector, based
on which the set of nodes are removed. GND repeats the
process until the end. Note that the unit-cost GND is just
the spectral cut method.
We use SNAP software1to implement the heuris-
tic methods, including Random, HDA, HBA, HCA and
HPRA. For the other baselines, we use the source
codes23456 released online, and use the defaut parameter
settings for each method.
Synthetic graphs
We evaluate all competitors against various synthetic
networks. Synthetic networks are the result of applying
generative function, present the advantage of displaying
specific topological features that are both a prior known
and tunable [29]. More specifically, we select a collection
of 4 most common network types, summarized in Table 1.
Note that there are many other random network models,
such as regular graphs, circle graphs, grid graphs, lad-
der graphs, etc, we do not consider them since they are
not difficult to dismantle, and there always exists some
effective heuristic methods for them.
1http://snap.stanford.edu/snap/
2https://github.com/zhfkt/ComplexCi
3http://power.itp.ac.cn/~zhouhj/codes.html
4https://github.com/abraunst/decycler
5https://github.com/hcmidt/corehd
6https://github.com/renxiaolong/Generalized-Network- Dismantling
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 5 of 13
Table 1 Overview of four random network types
Abbre Name Parameters
ER Erdos-Renyi n in [500, 800], p in [0.10,0.15,0.18,0.20,0.25],
WS Watt-Strogatz n in [500, 800], k in [5,6,7,8,9], p in
[0.1,0.2,0.3,0.4,0.5,0.6,0.7]
BA Barabasi-Albert n in [500, 800], m in [2,3,4,5,6]
PLC Powerlaw-Cluster n in [500, 800], m in [2,3,4,5,6], p in
[0.10,0.15,0.18,0.20,0.25]
Erdos-Renyi(ER) [8]. ER model is first introduced by
Paul Erdos and Alfred Renyi, it returns a Gn,pgraph, where
nis the graph nodes, pis the edge creation probability. The
Gn,pchooses each of the possible edges with probability
p. This model can be used in the probabilistic method to
prove the existence of graphs satisfying various properties,
or to provide a rigorous definition of what it means for a
property to hold for almost all graphs [8].
Watt-Strogatz(WS) [30]. WS is a random generative
model that produces graphs with small-world properties,
including short average path lengths and high clustering.
It was proposed by Duncan J. Watts and Steven Strogatz
in 1998. The tunable parameters include the node number
n,knearest neighbors in a ring topology that each node is
joined with, and the probability of rewiring each edge p.
Barabasi-Albert(BA) [3]. BA is a model that generates
random scale-free networks using a preferential attach-
ment mechanism. Many real-world networks are thought
to be approximately scale-free and contain few nodes
(called hubs) with unusually high degree as compared to
the other nodes. The BA model tries to explain the exis-
tence of such nodes in real networks. The algorithm is
named for its inventors Albert-Laszlo Barabasi and Reka
Albert and is a special case of a more general model called
Price’s model [28]. It generates a graph of nnodes by
attaching new nodes with each adding medges that are
preferentially attached to existing nodes with high degree.
Powerlaw-Cluster(PLC) [13].PLCisamodeforgen-
erating graphs with powerlaw degree distribution and
approximate average clustering. It is essentially the BA
growth model with an extra step that each random edge is
followed by a chance of making an edge to one of its neigh-
bors too (and thus a triangle) [13]. The model improves on
BA in the sense that it enables a higher average clustering
to be attained if desired. The tunable parameters include
thenumberofnodesn, the number of random edges to
add for each new node m, and the probability of adding a
triangle after adding a random edge p.
Figure 3visualizes one instance for each of the above
four networks types.
Real-world networks
We also conduct experiments on 18 real-world networks,
which cover a wide range of domains, including malicious
networks, PPI networks, infrastructure networks, social
networks, citation networks, communication networks,
etc. Specifically, they are:
Corruption [26], a malicious network where nodes are
people listed in scandals, and the ties indicate that two
people were involved in the same corruption scandal;
Crime [16], a malicious network from the projection
of a bipartite network of persons and crimes, each node
denotes a person, an edge represents that two person are
involved in the same crime;
USairport [16], a network of flights between US air-
ports in 2010. Each node is an airport, and each edge
represents a connection from one airport to another;
Hamster [16]. This Network contains friendships and
family links between users of the website hamster.com;
Figeys [16], a network of interactions between pro-
teins in Humans (Homo sapiens), from the first large-
scale study of protein–protein interactions in Human cells
using a mass spectrometry-based approach;
CA-GrQc [18], a collaboration network from the e-
print arXiv and covers scientific collaborations between
authors papers submitted to General Relativity and Quan-
tum Cosmology category;
HI-II-14, the corresponding Human Interactome
dataset covering Space II and reported in 2014.
Each node represents a distinct protein, each edge
denotes the interaction between the corresponding
proteins;
Fig. 3 Visualization of one instance for each type of random networks
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 6 of 13
Powergrid [16], a power grid network of the Western
States of the United States of America. An edge repre-
sents a power supply line. A node is either a generator, a
transformator or a substation;
CA-HepPh [18], a collaboration network from the e-
print arXiv and covers scientific collaborations between
authors papers submitted to High Energy Physics - Phe-
nomenology category;
DBLP [16], a citation network of DBLP, a database of
scientific publications such as papers and books. Each
node in the network is a publication, and each edge repre-
sents a citation of a publication by another publication;
Cora [16], a citation network of Cora. Nodes represent
scientific papers. An edge between two nodes indicates
that the left node cites the right node;
Digg [16], a reply network of the social news website
Digg. Each node in the network is a user of the website,
and each edge denotes that a user replied to another user;
Email-Enron [20], the Enron email communication net-
work which covers all the email communication within a
data set of around half million emails. Each node is an
email address, and an edge denotes at least one email
communication;
Brightkite [16], a social network contains user–user
friendship relations from Brightkite, a former location-
based social network were user shared their locations.
A node represents a user, and an edge indicates that a
friendship exists between the user represented by the left
node and the user represented by the right node;
Gnutella31 [19], a sequence of snapshots of the
Gnutella peer-to-peer file sharing network from August
2002. Nodes represent hosts in the Gnutella network
topology and edges represent connections between the
Gnutella hosts;
Facebook [16], contains friendship data of a small sub-
set of Facebook users. A node represents a user and an
edge represents a friendship between two users;
Epinion [16], the trust network from the online social
network Epinions. Nodes are users of Epinions and
directed edges represent trust between the users;
Douban [16], a social network of Douban, a Chinese
online recommendation site. A node represents a user of
Douban and an edge represents a friendship between two
users.
We treat all the networks as undirected ones and remove
the self-loops. We extract the largest connected com-
ponent. Basic statistics of the extracted networks are
reported as Table 2.
We also draw the degree distributions for these net-
works in Fig. 4.Wecanseemostrealnetworks(except
Corruption network) share an approximate scale-free
structure, which presents a well-known resilience against
random failures, but disintegrate rapidly under intentional
attacks targeting key nodes [2].
Table 2 Basic statistics for real-world networks. Ordered by the number of nodes. Values are for the giant component of the network.
MSP is the mean shortest path length, CC is the clustering coefficient, Assor is the assortativity, PE is the powerlaw exponent
Network N E MAX DEG AVG DEG Diameter MSP CC Assor PE Type
Corruption 309 3,281 86 21.24 7 2.99 0.9288 0.5324 5.13 Malicious
Crime 829 1,473 25 3.55 10 5.04 0.0058 –0.1645 2.65 Malicious
USairport 1,572 17,214 314 21.90 8 3.12 0.5048 -0.1134 3.07 Infrastructure
Hamster 2,000 16,098 273 16.10 10 3.59 0.5401 0.0227 2.67 Social
Figeys 2,217 6,418 314 5.79 10 3.84 0.0403 -0.3318 2.92 PPI
CA-GrQc 4,158 13,422 81 6.46 17 6.05 0.5569 0.6392 2.47 Collaboration
HI-II-14 4,165 13,087 286 6.28 11 4.16 0.0444 –0.2016 2.68 PPI
Powergrid 4,941 6,594 19 2.67 46 18.99 0.0801 0.0035 2.50 Infrastructure
CA-HepPh 11,204 117,619 491 21.00 13 4.67 0.6216 0.6295 2.67 Collaboration
DBLP 12,495 49,563 709 7.93 10 4.42 0.1178 -0.0461 3.22 Citation
Cora 23,166 89,157 377 7.70 20 5.74 0.2660 -0.0553 2.44 Citation
Digg 29,652 84,781 310 5.72 12 4.68 0.0054 0.0027 2.46 Communication
Email-Enron 33,696 180,811 1,383 10.73 11 4.03 0.5092 –0.1165 2.62 Communication
Brightkite 56,739 212,945 1,134 7.51 18 4.86 0.1734 0.0096 2.50 Social
Gnutella31 62,561 147,878 95 4.73 11 5.96 0.0055 –0.0927 2.91 Infrastructure
Facebook 63,392 816,831 1,098 25.77 15 4.31 0.2218 0.1768 2.92 Social
Epinions 75,877 405,739 3,044 10.69 15 4.40 0.1378 –0.0406 2.69 Social
Douban 154,908 327,162 287 4.22 9 5.10 0.0161 -0.1803 2.65 Social
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 7 of 13
Fig. 4 Degree distribution of the real-world networks. X axis is the node degree, y axis is the number of nodes corresponding to that degree, and we
choose double logarithmic coordinates to more intuitively determine whether it is scale-free
Results
In this section, we first demonstrate the effectiveness of
the reinsertion technique on both synthetic graphs and
real-world networks, then we explore the effects of differ-
ent reinsertion techniques.
Synthetic results
We test all methods w/o the reinsertion technique on
synthetic graphs randomly generated by four classic mod-
els introduced in “Synthetic graphs” section. For each
model, we generate 100 graphs with the parameters in
Table 3 Comparison results (%) on synthetic graphs without reinsertion. Each result is averaged over 100 test instances. The result
format is mean±variance. The bold ones indicate the best results for that network
R(No reinsert) ER WS BA PLC Avg
Random 49.91±0.02 44.19±3.66 45.80±2.87 46.37±2.66 46.57±2.30
HDA 49.55±0.25 32.44±5.32 21.43±7.70 21.89±7.53 31.33±5.20
HBA 49.47±0.29 25.55±7.59 19.96±7.65 20.23±7.52 28.80±5.76
HCA 49.58±0.23 28.33±7.07 20.99±7.87 21.26±7.72 30.04±5.72
HPRA 49.55±0.25 33.10±4.88 22.46±7.77 22.92±7.64 32.01±5.13
CI 49.70±0.25 31.21±5.81 21.94±7.67 22.35±7.51 31.30±5.31
MinSum 49.45±0.31 32.43±5.09 20.58±7.70 21.40±7.36 30.96±5.12
BPD 49.58±0.24 33.69±4.60 22.75±7.63 23.21±7.49 32.31±4.99
CoreHD 49.56±0.26 32.48±5.33 21.01±7.76 21.72±7.50 31.19±5.21
GND 49.62±0.14 27.10±7.62 22.89±7.67 23.01±7.42 30.66±5.71
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 8 of 13
Table 4 Comparison results (%) on synthetic graphs with reinsertion. Each result is averaged over 100 test instances. The result format
is mean±variance. The bold ones indicate the best results for that network
R(With reinsert) ER WS BA PLC Avg
Random 48.97±0.44 33.12±4.24 27.24±7.17 28.14±7.56 34.37±4.85
HDA 48.84±0.51 31.36±4.59 20.75±7.20 21.21±7.12 30.54±4.86
HBA 48.92±0.46 33.18±4.28 25.41±7.47 25.90±7.79 33.35±5.00
HCA 48.87±0.48 31.72±4.43 20.90±7.24 21.27±7.13 30.69±4.82
HPRA 48.84±0.51 31.32±4.61 20.77±7.22 21.16±7.13 30.52±4.87
CI 49.67±0.29 31.16±5.47 21.68±7.53 22.12±7.41 31.16±5.17
MinSum 48.80±0.53 31.61±4.48 20.78±7.22 21.19±7.10 30.59±4.83
BPD 49.39±0.35 31.22±4.97 21.17±7.45 21.60±7.34 30.84±5.03
CoreHD 49.36±0.36 31.48±5.02 21.24±7.42 21.66±7.35 30.93±5.04
GND 49.20±0.33 31.84±5.67 21.67±7.62 22.10±7.60 31.20±5.30
Table 5 The promotion of R(%) on synthetic graphs with reinsertion. Each result is averaged over 100 test instances. The result format
is mean±variance. The bold ones indicate the best results for that network
Promotion ER WS BA PLC Avg
Random 1.88±0.86 24.96±8.41 41.23±12.72 40.01±13.75 27.02±8.94
HDA 1.43±0.53 2.77±7.41 2.54±2.57 2.67±2.27 2.35±3.19
HBA 1.13±0.36 -43.55±54.37 -33.61±19.64 -33.95±21.25 -27.50±23.91
HCA 1.44±0.53 -18.16±30.91 -1.07±4.68 -1.73±5.42 -4.88±10.39
HPRA 1.45±0.55 5.12±7.94 7.50±2.09 7.76±2.09 5.46±3.17
CI 0.05±0.13 -0.11±2.34 1.08±0.92 0.96±0.84 0.49±1.06
MinSum 1.33±0.46 2.16±4.85 -2.38±4.29 0.52±3.16 0.41±3.19
BPD 0.40±0.26 7.62±3.18 7.59±2.70 7.71±3.35 5.83±2.37
CoreHD 0.41±0.28 2.95±3.00 -2.18±3.55 -0.07±2.38 0.28±2.30
GND 0.85±0.52 -22.46±21.35 5.97±4.67 4.96±4.20 -2.67±7.68
Table 6 Time (/s) comparison of different methods on synthetic graphs. Each result is averaged over 100 test instances. The result
format is mean±variance. The bold ones indicate the best results for that network
Time/s ER WS BA PLC Avg
HDA 0.01±0.00 0.00±0.00 0.00±0.00 0.00±0.00 0.00±0.00
HBA 426.71±271.03 41.03±20.11 31.53±18.58 33.42±20.92 133.17±82.66
HCA 311.23±201.09 20.21±9.83 14.10±8.49 14.83±9.47 90.09±57.22
HPRA 0.77±0.24 0.53±0.13 0.45±0.12 0.46±0.13 0.55±0.15
CI 8.67±6.30 0.02±0.01 0.03±0.03 0.03±0.02 2.19±1.59
MinSum 22.21±9.92 2.28±0.47 1.93±0.81 2.69±0.80 7.28±3.00
BPD 148.21±100.97 0.85±0.46 0.94±0.66 1.02±0.68 37.76±25.69
CoreHD 3.06±1.14 0.12±0.04 0.18±0.08 0.20±0.09 0.89±0.34
GND 0.86±0.35 0.18±0.10 0.15±0.08 0.14±0.09 0.33±0.15
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 9 of 13
Table 7 Comparison results (%) on real-world networks without reinsertion. The bold result is the best one of that network
R(No reinsert) Random HDA HPRA CI MinSum BPD CoreHD GND
Corruption 38.50 8.48 8.81 16.42 33.27 33.74 8.37 6.37
Crime 38.64 11.51 11.83 36.77 17.95 36.59 36.74 37.61
USairport 43.73 11.39 10.15 13.02 13.12 13.01 12.44 8.94
Hamster 44.40 19.00 16.02 16.36 23.27 22.77 20.23 15.58
Figeys 37.93 3.13 3.16 3.38 4.27 4.46 3.65 3.90
CA-GrQc 37.41 10.81 8.56 9.37 12.30 12.38 11.33 7.48
HI-II-14 40.60 5.75 5.89 22.41 7.28 23.14 23.18 23.80
Powergrid 21.47 5.25 5.90 5.23 5.58 4.74 5.49 2.20
CA-HepPh 42.41 18.55 14.78 16.71 20.84 20.71 19.31 13.14
DBLP 41.32 10.65 10.00 11.16 12.29 12.21 11.98 11.07
Cora 42.87 14.85 14.92 14.41 16.82 16.65 15.36 10.98
Digg 40.16 8.75 9.12 29.00 26.39 26.35 26.53 27.72
Email-Enron 39.63 4.53 4.18 18.97 6.30 19.20 19.88 16.78
Brightkite 39.30 8.73 8.79 8.47 9.55 22.07 9.03 9.05
Gnutella31 39.66 11.47 10.88 27.24 11.74 26.44 26.62 26.95
Facebook 45.81 27.24 27.06 41.68 27.91 41.22 41.58 41.91
Epinions 38.61 5.19 5.20 18.26 6.11 18.26 18.47 18.39
Douban 36.73 2.38 2.20 2.76 2.89 2.87 2.93 2.63
Avg 39.40 10.43 9.86 17.31 14.33 19.82 17.40 15.81
Table 8 Comparison results (%) on real-world networks with reinsertion. The bold result is the best one of that network
R(With reinsert) Random HDA HPRA CI MinSum BPD CoreHD GND
Corruption 12.78 11.66 11.73 18.40 18.63 18.34 18.82 21.14
Crime 14.19 11.33 11.41 36.47 14.18 36.32 36.12 36.60
USairport 14.49 9.85 9.76 11.68 9.61 9.58 9.65 9.82
Hamster 18.31 14.71 15.16 15.57 16.23 16.38 16.66 16.24
Figeys 3.99 3.14 3.13 3.28 3.44 3.39 3.43 3.76
CA-GrQc 10.31 8.59 8.66 8.70 8.85 8.67 8.75 7.70
HI-II-14 7.00 5.68 5.70 22.00 5.76 22.75 22.76 23.07
Powergrid 7.93 6.94 7.16 4.90 5.02 4.92 5.08 4.44
CA-HepPh 15.40 14.41 14.49 14.53 15.39 15.14 15.19 13.98
DBLP 10.42 8.75 8.75 8.94 10.18 10.05 10.15 9.79
Cora 17.00 13.48 13.49 13.11 14.05 13.79 14.04 13.83
Digg 10.96 8.59 8.60 28.21 26.28 26.24 26.24 26.94
Email-Enron 6.01 3.96 3.95 17.91 4.25 18.46 18.68 16.20
Brightkite 10.18 8.20 8.21 7.69 8.41 22.06 8.35 8.41
Gnutella31 11.20 10.07 10.07 25.98 10.83 25.87 25.86 25.46
Facebook 25.38 22.39 22.37 40.67 25.65 40.67 40.68 39.96
Epinions 5.70 4.96 4.94 17.95 4.97 18.18 18.18 18.00
Douban 2.72 2.09 2.09 2.17 2.38 2.36 2.36 2.31
Avg 11.33 9.38 9.43 16.56 11.34 17.40 16.72 16.54
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 10 of 13
Table 9 The promotion of R(%) on real-world networks with reinsertion. The bold result is the best one of that network
Promotion of RRandom HDA HPRA CI MinSum BPD CoreHD GND
Corruption 66.81 -37.50 -33.14 -12.06 44.00 45.64 -124.85 -231.87
Crime 63.28 1.56 3.55 0.82 21.00 0.74 1.69 2.69
USairport 66.86 13.52 3.84 10.29 26.75 26.36 22.43 -9.84
Hamster 58.76 22.58 5.37 4.83 30.25 28.06 17.65 -4.24
Figeys 89.48 -0.32 0.95 2.96 19.44 23.99 6.03 3.59
CA-GrQc 72.44 20.54 -1.17 7.15 28.05 29.97 22.77 -2.94
HI-II-14 82.76 1.22 3.23 1.83 20.88 1.69 1.81 3.07
Powergrid 63.06 -32.19 -21.36 6.31 10.04 -3.80 7.47 -101.82
CA-HepPh 63.69 22.32 1.96 13.05 26.15 26.90 21.34 -6.39
DBLP 74.78 17.84 12.50 19.89 17.17 17.69 15.28 11.56
Cora 60.35 9.23 9.58 9.02 16.47 17.18 8.59 -25.96
Digg 72.71 1.83 5.70 2.72 0.42 0.42 1.09 2.81
Email-Enron 84.83 12.58 5.50 5.59 32.54 3.85 6.04 3.46
Brightkite 74.10 6.07 6.60 9.21 11.94 0.05 7.53 7.07
Gnutella31 71.76 12.21 7.44 4.63 7.75 2.16 2.85 5.53
Facebook 44.60 17.80 17.33 2.42 8.10 1.33 2.16 4.65
Epinions 85.24 4.43 5.00 1.70 18.66 0.44 1.57 2.12
Douban 92.59 12.18 5.00 21.38 17.65 17.77 19.45 12.17
Avg 71.56 5.88 2.11 6.21 19.85 13.36 2.27 -18.02
Tabl e 1,andreportthevaluesofmeanandstandard
variance results. Table 3shows the comparison results
of Eq. 1without reinsertion, we can clearly see that
HBA the best across different types of networks, which
is widely validated by previous research [14,27], since
Table 10 Time (/s) comparison of different methods on
real-world networks. The bold result is the best one of that
network
Time/s HDA HPRA CI MinSum BPD CoreHD GND
Corruption 0.00 0.11 0.02 3.13 2.00 0.08 0.18
Crime 0.00 0.55 0.01 1.47 0.30 0.09 0.11
USairport 0.01 1.96 0.61 13.20 17.00 1.20 0.55
Hamster 0.01 4.74 0.44 13.89 7.00 0.90 0.68
Figeys 0.00 1.57 0.06 5.06 2.00 0.41 0.13
CA-GrQc 0.04 18.33 0.07 12.63 3.00 0.46 1.29
HI-II-14 0.02 10.66 0.49 10.51 9.00 1.98 1.32
Powergrid 0.05 23.06 0.01 5.90 1.00 0.21 1.48
CA-HepPh 0.31 170.96 13.90 138.90 140.00 8.50 20.69
DBLP 0.13 97.31 2.81 46.97 21.00 3.33 6.30
Cora 1.07 593.75 2.12 97.97 22.00 4.91 189.27
Digg 1.12 702.55 10.99 103.33 31.00 11.52 55.01
Email-Enron 1.63 876.55 70.74 242.27 86.00 36.20 35.70
Brightkite 4.47 2974.39 32.63 298.20 4.00 20.13 181.76
Gnutella31 3.05 2509.09 9.38 172.59 41.00 11.87 386.31
Facebook 9.03 6706.60 2723.23 1044.96 839.00 214.83 4614.54
Epinions 5.76 4115.42 668.43 522.51 254.00 142.92 233.14
Douban 2.88 3832.00 87.38 337.04 79.00 30.20 186.03
Avg 1.64 1257.76 201.30 170.58 86.57 27.21 328.58
HBA adaptively removes the highest betweenness nodes,
which are key to the whole network connectivity. HCA,
which adaptively removes the highest closeness nodes,
also performs excellently due to the similar reasons. How-
ever, considering the high computational costs of these
two methods (Table 6), they are not practical in large or
even medium scale networks. We can also see in methods
achieve good results in ER graphs, since these graphs are
purely random ones that there are no ’critical’ nodes that
determine the graph connectivity.
In Table 4, we enhance each method with the reinsertion
technique introduced in “Reinsertion technique”section,
and report the refined results, and we also show the
Table 11 Average promotion of Ron synthetic graphs for
different reinsertion techniques. Each result is averaged over all
test graphs (including four types of graphs, and 100 graphs for
each type), and the result format is mean±variance. The bold
result is the best one for that method
Avg Promotion of RReinsert_I Reinsert_II Reinsert_III
Random 27.02±8.94 32.73±10.87 33.36±10.74
HDA 2.35±3.19 9.96±5.16 10.67±4.81
HBA -27.50±23.91 -11.47±7.94 -10.53±7.82
HCA -4.88±10.39 5.08±1.80 5.82±1.81
HPRA 5.46±3.17 12.68±5.68 13.48±5.43
CI 0.49±1.06 -0.52±1.07 -0.14±0.93
MinSum 0.41±3.19 8.17±5.41 8.86±5.11
BPD 5.83±2.37 4.73±2.95 5.86±3.34
CoreHD 0.28±2.30 -0.96±2.81 0.35±2.77
GND -2.67±7.68 -4.50±7.57 -3.04±7.11
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 11 of 13
promotion (Eq. 2) after adding the reinsertion in Table 5.
We can see that most methods (except for HBA, HCA
and GND) get improved after using reinsertion, and on
average, HPRA (reinserted) performs the best among all.
We also observe two interesting things: i) The best per-
formed HBA gets deteriorated greatly when utilized with
reinsertion, however, even the best result for reinserted
methods (HPRA) cannot beat the vanilla HBA (Table 3).
This indicates that the vanilla HBA has achieved the
close-to-optimal performance for the network disman-
tling problem, at which the reinsertion is no longer a
refinement, but a hindrance; ii) The pure Random strat-
egy gets greatly improved with reinsertion, making the
reinserted random strategy be close to those manually-
designed state-of-the-arts.
promotion =(Roriginal −Rreinsert )
Roriginal
(2)
However, if taking account of running time, we find
actually the simple heuristic HDA achieves the best bal-
ance between effectiveness and efficiency (Tables 3,4and
6). The reinserted HDA is only 1.74% worse than the best
result (vanilla HBA), while is hundreds of times faster
(Table 6). Note that we do not list the time for Random
strategy, since it basically takes no time to obtain a random
solution.
Real-world results
Now we will see the effects of reinsertion on real-world
networks. Since HBA and HCA are computationally pro-
hibitive on medium or large networks (e.g., HBA takes
over 5 days to finish computation on the Cora network,
with 23,166 nodes and 89,157 edges.), we do not compare
with them in this section.
Tabl e 7shows the results of vanilla methods without
reinsertion. We can see that HDA, HPRA and GND per-
forms relatively better than other methods, and HPRA
is the best (0.0986) among all, and followed by HDA
(0.1043). Table 8gives the results after reinsertion, and
Tabl e 9shows the promotion results. Consistent with the
observations from synthetic results, most methods get
improvements for different levels, with the refinement of
reinsertion. For example, the random strategy obtains an
average 71.56% gain (Table 9) with reinsertion, making it
even beat the state-of-the-art MinSum strategy (Table 8).
Among the reinserted methods, HDA achieves the highest
performance with an average 0.0938 (Table 8) robustness
score (Eq. 1). However, GND is deteriorated on some net-
works when refined with reinsertion (Table 8), the reason
behind remains to be explored. When considering the exe-
cution, HDA is far more efficient than the other ones, e.g.,
it is about 767 times faster than HPRA, which is very close
to HDA in effectiveness.
Effects of different reinsertion strategies
We have observed the impressive gains brought by
the reinsertion technique in “Synthetic results”and
“Real-world results” sections, now we may ask: Is the rein-
sertion in “Reinsertion technique” section the best one?
Does there exist more effective reinsertion methods? In this
section, we try to answer this question by exploring other
potential reinsertion techniques (Table 10).
We name the previous reinsert method as Reinsert_I,
and here we propose two other ones, and call them
Reinsert_II and Reinsert_III respectively. Basically, the
general reinsertion technique is to add back one of the
removed node (together with the adjacent edges), cho-
sen based on some criteria, until all nodes are back in
the network. Different reinsertion methods define differ-
ent criteria, based on which, we define the following three
reinsertion strategies:
•Reinsert_I: The
criteria
is once reinserted, it joins
the smallest number of clusters;
•Reinsert_II: The
criteria
is once reinserted, it joins
the clusters of smallest sizes;
•Reinsert_III: The
criteria
is once reinserted, it joins
the clusters minimizing the multiply of both numbers
and sizes;
In Fig. 2, each node is assigned an index c(i)given by the
criteria specified by the reinsertion technique. For Rein-
sert_I,c(red)=2, c(blue)=4, c(green)=3, then the red
node is reinserted; for Reinsert_II,c(red)=10, c(blue)=
5, c(green)=6, then the blue node is reinserted; for
Reinsert_III,c(red)=20, c(blue)=20, c(green)=18,
then the green node is reinserted. After that, the c(i)s
are recalculated and the new node with smallest c(i)is
found and reinserted. Repeat these steps until the end. As
a consequence, different reinsertion strategies determines
different nodes to be reinserted first, leading to differ-
ent refinement results. To decide which one is better in
Table 12 Average promotion of Ron real-world networks for
different reinsertion techniques. Each result is averaged over all
test networks (total 18 real-world networks), and the bold ones
are the best results for that method
Avg Promotion of RReinsert_I Reinsert_II Reinsert_III
Random 71.56 77.31 78.24
HDA 5.88 21.71 23.98
HPRA 2.11 19.14 21.45
CI 6.21 4.46 6.01
MinSum 19.85 16.72 20.59
BPD 13.36 12.33 14.49
CoreHD 2.27 6.92 9.79
GND -18.02 -9.43 -8.16
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 12 of 13
practice, we compare the average performance promotion
for each method on both synthetic graphs and real-world
networks (Tables 11 and 12).
It can be clearly observed in Tables 11 and 12 that Rein-
sert_III achieves the most promotions for most methods
(except CI) on both synthetic and real-world networks,
compared to other two reinsertion strategies, and excels
to a significant extent to the current strategy Reinsert_I.
For CI method, Reinsert_I tendstobemoreeffective.All
the three reinsertion strategies fail in HBA and GND.
To illustrate the effects of these three strategies more
intuitively, we draw the robustness curve of CA-GrQc net-
work for different methods with different reinsertions in
Fig. 5, which is plotted with horizontal axis being the
fraction of removed nodes, and vertical axis being the
remaining giant connected component size. Actually, the
value of Eq. 1approximates the area under the robust-
ness curve. The figure clearly shows that the reinsertion
greatly helps reduce the area under the curve, compared to
the original method, and Reinsert_III is among the most
effective one, while all the reinsertions produce negative
effects on the GND method.
Conclusion
In this paper, we, for the first time, systematically explore
the effects of reinsertion techniques for the network dis-
mantling problem. Previous research tend to use their
reinserted results to compare with other un-reinserted
baseline methods, which may mislead us in the selection
of the real best dismantling strategy for applications at
hand. We conduct comprehensive ablation studies on
both synthetic graphs generated by four classical ran-
dom network models, i.e., ER, WS, BA and PLC, and
18 real-world networks across seven different domains
and with different scales, and the results show that: i)
HBA (no reinsertion) is the most effective network dis-
mantling strategy, however, it can only be applicable in
small scale networks; ii) HDA (with reinsertion) achieves
the best balance between effectiveness and efficiency. It
is surprising that such a simple heuristic method would
beat most state-of-the-art methods if enhanced with rein-
sertion techniques; iii) The reinsertion technique helps
improve the performance for most current methods,
except for HBA, HCA and GND (on small-world type
graphs); iv) Reinsert_III, which determines the node based
on that it joins the clusters minimizing the multiply of
both numbers and sizes, is the most effective reinsertion
strategy for most methods (except for CI, where Rein-
sert_I suits best). We believe the results in this paper could
provide as a reference for choosing and designing the
most effective strategy for realistic network dismantling
applications.
However, we still lack a deep understanding about why
such a simple reinsertion technique works so well for
the network dismantling problem, which would be a very
meaningful future research topic to be explored. We will
later release the codes and data to support the research in
this direction.
Fig. 5 Robustness curve of CA-GrQc for different reinsertion techniques
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
Fan et al. Journal of Cloud Computing: Advances, Systems and Applications (2020) 9:24 Page 13 of 13
Acknowledgments
We thank Yuan Liu for her valuable discussions.
Authors’ contributions
CF and YF initiated the project. CF and YF, XB designed and managed the
project. CF and LZ performed calculations. All authors analyzed the results,
wrote the manuscript, and edited the manuscript. All author(s) read and
approved the final manuscript.
Authors’ information
CF, LZ, YF, JH and ZL are all affiliated with College of Systems Engineering,
National University of Defense Technology. CF and LZ are both Ph.D.
candidates. YF is an associate professor, JH and ZL are both professors. XB is a
professor in School of Systems Science and Engineering, Sun Yat-sen
University.
Funding
This work was supported by CSC scholarship offered by Chinese Scholarship
Council and NSFC-71701205.
Availability of data and materials
Upon reasonable requests, all code/data used in the analysis will be available
to any researcher for purposes of reproducing or extending the analysis.
Competing interests
The authors declare that they have no competing interests.
Author details
1College of Systems Engineering, National University of Defense Technology,
Changsha, Hunan, China. 2School of Systems Science and Engineering, Sun
Yat-sen University, Guangzhou, Guangdong Province, China.
Received: 30 June 2019 Accepted: 3 April 2020
References
1. Albert R, Barabási AL (2002) Statistical mechanics of complex networks.
Rev Mod Phys 74(1):47
2. Albert R, Jeong H, Barabási AL (2000) Error and attack tolerance of
complex networks. Nature 406(6794):378
3. Barabási AL, Albert R (1999) Emergence of scaling in random networks.
Science 286(5439):509–512
4. Bavelas A (1950) Communication patterns in task-oriented groups. J
Acoust Soc Am 22(6):725–730
5. Braunstein A, Dall’Asta L, Semerjian G, Zdeborová L (2016) Network
dismantling. Proc Natl Acad Sci 113(44):12,368–12,373
6. Brin S, Page L (1998) The anatomy of a large-scale hypertextual web
search engine. Comput Netw ISDN Syst 30(1-7):107–117
7. Cohen R, Erez K, Ben-Avraham D, Havlin S (2001) Breakdown of the
internet under intentional attack. Phys Rev Lett 86(16):3682
8. ERDdS P R&wi (1959) On random graphs i. Publ Math Debrecen 6:290–297
9. Fan C, Xiao K, Xiu B, Lv G (2014) A fuzzy clustering algorithm to detect
criminals without prior information. In: Proceedings of the 2014 IEEE/ACM
International Conference on Advances in Social Networks Analysis and
Mining. IEEE Press. pp 238–243
10. Fan C, Liu Z, Lu X, Xiu B, Chen Q (2017) An efficient link prediction index
for complex military organization. Phys A Stat Mech Appl 469:572–587
11. Fan C, Zeng L, Ding Y, Chen M, Sun Y, Liu Z (2019) Learning to identify
high betweenness centrality nodes from scratch: A novel graph neural
network approach. In: Proceedings of the 28th ACM International
Conference on Information and Knowledge Management. pp 559–568
12. Freeman LC (1977) A set of measures of centrality based on betweenness.
Sociometry: 35–41
13. Holme P, Kim BJ (2002) Growing scale-free networks with tunable
clustering. Phys Rev E 65(2):026,107
14. Holme P, Kim BJ, Yoon CN, Han SK (2002) Attack vulnerability of complex
networks. Phys Rev E 65(5):056,109
15. Kempe D, Kleinberg J, Tardos É (2003) Maximizing the spread of influence
through a social network. In: Proceedings of the ninth ACM SIGKDD
international conference on Knowledge discovery and data mining. ACM.
pp 137–146
16. Kunegis J (2013) Konect: the koblenz network collection. In: Proceedings
of the 22nd International Conference on World Wide Web. ACM.
pp 1343–1350
17. Lalou M, Tahraoui MA, Kheddouci H (2018) The critical node detection
problem in networks: a survey. Comput Sci Rev 28:92–117
18. Leskovec J, Krevl A (2014) SNAP Datasets: Stanford large network dataset
collection. http://snap.stanford.edu/data
19. Leskovec J, Kleinberg J, Faloutsos C (2007) Graph evolution: Densification
and shrinking diameters. ACM Trans Knowl Discov Data (TKDD) 1(1):2
20. Leskovec J, Lang KJ, Dasgupta A, Mahoney MW (2009) Community
structure in large networks: Natural cluster sizes and the absence of large
well-defined clusters. Internet Math 6(1):29–123
21. Morone F, Makse HA (2015) Influence maximization in complex networks
through optimal percolation. Nature 524(7563):65
22. Mugisha S, Zhou HJ (2016) Identifying optimal targets of network attack
by belief propagation. Phys Rev E 94(1):012,305
23. Pastor-Satorras R, Vespignani A (2001) Epidemic spreading in scale-free
networks. Phys Rev Lett 86(14):3200
24. Ren XL, Gleinig N, Helbing D, Antulov-Fantulin N (2018) Generalized
network dismantling. arXiv preprint arXiv:180101357
25. Ren XL, Gleinig N, Helbing D, Antulov-Fantulin N (2019) Generalized
network dismantling. Proc Natl Acade Sci 116(14):6554–6559
26. Ribeiro HV, Alves LG, Martins AF, Lenzi EK, Perc M (2018) The dynamical
structure of political corruption networks. J Complex Netw 6(6):989–1003
27. Schneider CM, Mihaljev T, Herrmann HJ (2012) Inverse targeting—an
effective immunization strategy. EPL (Europhys Lett) 98(4):46,002
28. Von Luxburg U (2007) A tutorial on spectral clustering. Stat Comput
17(4):395–416
29. Wandelt S, Sun X, Feng D, Zanin M, Havlin S (2018) A comparative analysis
of approaches to network-dismantling. Sci Rep 8(1):13,513
30. Watts DJ, Strogatz SH (1998) Collective dynamics of ’small-world’
networks. Nature 393(6684):440
31. Zdeborová L, Zhang P, Zhou HJ (2016) Fast and simple decycling and
dismantling of networks. Sci Rep 6:37,954
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY 4.0
Content may be subject to copyright.