Title: Positive and Negative Relationships (Chapter 5)
Book: Networks, Crowds, and Markets
Reasoning about a Highly Connected World
Presenter: Saeid Ghasemshirazi
7. Android Malware Adventures
Structural Balance Property
7
Networks, Crowds, Markets: Reasoning about a Highly Connected World
For every set of three nodes, if we consider the
three edges connecting them, either all three of
these edges are labeled +, or else exactly one of
them is labeled +
8. Android Malware Adventures
Balance theorem
8
Networks, Crowds, Markets: Reasoning about a Highly Connected World
If a complete graph is balanced then either all pairs of nodes are friends or else the
nodes can be divided into two groups, X and Y, such that:
The people in X all like one another, likewise in Y
Everyone in X is the enemy of everyone in Y
9. Android Malware Adventures
Proof of the Balance Theorem, arbitrary A
9
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Recall:
Every two nodes in X are friends.
Every two nodes in Y are friends.
Every node in X is an enemy of every node in Y .
10. Android Malware Adventures
Applications of Structural Balance
10
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Use for Friendships
--------------------
Use for Antagonist
the nations GB, Fr, Ru, It, Ge,and AH are Great Britain, France, Russia, Italy, Germany, and Austria-Hungary respectively
12. Android Malware Adventures
Weak Structured Balanced Networks
12
Networks, Crowds, Markets: Reasoning about a Highly Connected World
There is no set of three nodes
such that the edges amongst
them consist of two positive
edges and one negative edge
We allow a triangle that is all
negative, unlike structured
balanced networks
Characterization theorem:
If a labeled complete graph is
weakly balanced then it’s nodes
can be divided into groups that
that every two nodes in the same
group are friends and every two
nodes in different groups are
enemies
13. Android Malware Adventures
Allowing negative triangles : Weakly Balanced Networks
13
Networks, Crowds, Markets: Reasoning about a Highly Connected World
15. Android Malware Adventures
Structural Balance in Non-Complete(Arbitrary) Networks
15
Networks, Crowds, Markets: Reasoning about a Highly Connected World
16. Android Malware Adventures
Structural Balance in Non-Complete Networks (Disallowing negative cycles)
16
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Two equivalent definitions:
Local view :
a (non-complete) graph is balanced if it can be “completed” by adding edges to form a signed complete graph that is
balanced .
Global view :
we could define a signed graph to be balanced if it is possible to divide the nodes into two sets X and Y ,such that any
edge with both ends inside X or both ends inside Y is positive, and any edge with one end in X and the other in Y is
negative
17. Android Malware Adventures
Structural Balance in Non-Complete Networks (Disallowing negative cycles)
17
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Two equivalent definitions
18. Android Malware Adventures
Characterization
18
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Theorem:
A signed graph is balanced if and only if it contains no cycle with an odd number(2k+1) of negative edges
19. Android Malware Adventures
Proof(Create Positive Supernodes)
19
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Convert the graph to a reduced one in which there are only negative edges
Reduced graph
20. Android Malware Adventures
Negative cycle in supernode graph gives negative cycle in original graph
20
Networks, Crowds, Markets: Reasoning about a Highly Connected World
21. Android Malware Adventures
Negative cycle in supernode graph gives negative cycle in original graph
21
Networks, Crowds, Markets: Reasoning about a Highly Connected World
22. Android Malware Adventures
BFS of the Reduces Graph
22
Networks, Crowds, Markets: Reasoning about a Highly Connected World
The BFS has a cross edge if the graph is not bipartite
If the graph is bipartite then we can add even(2k) layers to X,
odd(2k+1) to Y, and all negative edges go between X and Y
If the graph is not bipartite, there is a negative odd cycle (2k+1)
A more standard drawing of the reduced graph
node G as the starting root node
The D-A,D-B path have the same length k
23. Android Malware Adventures
What if we only know that most triangles are balanced?
23
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Claim:
If at least 99.9% of all triangles in a labeled complete graph are balanced, then either
• there is a set consisting of at least 90% of the nodes in which at least 90% of all pairs are friends, or
else.
• the nodes can be divided into two groups, X and Y , such that:
(i)at least 90% of the pairs in X like each other.
(ii) at least 90% of the pairs in Y like each other.
(iii) at least 90% of the pairs with one end in X and the other end in Y are enemies.
.
24. Android Malware Adventures
Approximately Balanced Networks(General Statement)
24
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Theorem:
Let 𝜖 be any real 0 ≤ 𝜖 < 1/8, and let 𝛿 = 𝜖
1
3.
If at least 1 − 𝜖 of the triangles in a complete labeled graph are balanced then either
There is a set consisting of 1 − 𝛿 of the nodes where at least 1 − 𝛿 of the pairs are
friends, or else
The nodes can be divided into two sets X,Y, such that
At least 1 − 𝛿 of the pairs in X like each other
At least 1 − 𝛿 of the pairs in Y like each other
At least 1 − 𝛿 of the pairs with one edge in X and the other in Y are enemies.
In Balance Theorem: 𝜖=0 , 𝛿=0
In Approximately Balance Network:𝜖=0.001 , 𝛿=0.1
25. Android Malware Adventures
Approximately Balanced Networks Proof
25
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Assume graph is balanced, show it has one of the two forms
• As before, define two sets X and Y to be friends and enemies of some node A
• A must be a “good” node: a node that isn’t involved in too many unbalanced triangles
• Given a complete graph with N nodes
How many edges?
N(N-1)/2
How many triangles?
N(N-1)(N-2)/6
Triangles with A,B,C : ABC; ACB; BAC; BCA; CAB; and CBA
26. Android Malware Adventures
Approximately Balanced Networks Proof
26
Networks, Crowds, Markets: Reasoning about a Highly Connected World
There are at most ε N(N-1)(N-2)/6 unbalanced triangles
• Find each node’s weight: the number of unbalanced triangles it participates in
• Total weight of all nodes : 3N(N − 1)(N − 2)/6 =N(N − 1)(N − 2)/2
• Average weight of all nodes : N(N − 1)(N − 2)/2 *1/N=ε (N-1)(N-2)/2
• Pick some node A with weight less than the average They actually use at most ε N^2/2
• Since A is involved in less than ε N^2/2 unbalanced triangles, properties
within and between X and Y fail < ε N^2/2 times .
• Determine if one of X or Y is “small”
If so, most nodes are in a single mutually friendly group
If not, the nodes are divided into two mutually friendly groups who are mutual enemies
27. Android Malware Adventures
Approximately Balanced Networks Proof
27
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Let x =|X| be the number of nodes in X and y = |Y| be the number of nodes in Y
• Suppose first that X is big, x≥(1-δ)N
•
X has x(x-1)/2 internal edges
•
X contains at least (1-δ)N of the nodes
• At least (1-δ)N of the pairs in X are friends
• Same arguments hold if Y is big
28. Android Malware Adventures
Approximately Balanced Networks Proof
28
Networks, Crowds, Markets: Reasoning about a Highly Connected World
Suppose neither X nor Y is big:
x<(1-δ)N
y<(1-δ)N
• There are XY edges between, at least xy≤δ N^2/2
• There are at most ε N2/2 positive edges with one end in X and the other
in Y , so as a fraction of the total this is at most:
• There are x(x-1)/2 internal edges in X (same for Y)
• Since x>δN, this total is at most δ^2 N^2/2 (internal edges)
• There are at most ε N2/2 negative edges with one end in X and the other
in Y , so as a fraction of the total this is at most :