Adversarial Risk Analysis for Urban Security

 
Adversarial Risk Analysis for
Urban Security
 
Jesus Rios, IBM Research
David Rios, Cesar Gil, D. Banks
 
Outline
 
Adversarial Risk Analysis
The sequential Defend-Attack-Defend Model
 
Urban security resource allocation problem
Discussion
 
2
 
Adversarial Risk Analysis
 
A framework to manage risks from actions of intelligent
adversaries
 
One-sided prescriptive support
Use a SEU model
Treat the adversary
s decision as uncertainties
 
New method to predict adversary
s actions
We assume the adversary is an 
expected utility maximizer
Model his decision problem
Assess his probabilities and utilities
Find his action of maximum expected utility
But other 
descriptive
 models are possible
 
Uncertainty in the Attacker’s 
decision
 stems from
our
 uncertainty about his 
probabilities and utilities
 
3
 
Defend–Attack–Defend model
 
Two intelligent players
Defender
Attacker
 
Sequential moves
Defender moves first, deploying defense resources
Attacker observes Defender’s move and choose an attack
Defender moves again responding to the attack
 
 
 
 
4
 
Defend–Attack–Defend model
 
5
 
Standard Game Theory Analysis
 
Under common knowledge of utilities and probs
At node
 
 
Expected utilities at node S
 
 
 
 
Best Attacker
s decision at node A
 
 
Best Defender
s decision at node
 
 
Solution:
 
6
 
ID reduction algorithm for solving the
sequential D-A-D game
 
7
 
ARA: Supporting the Defender
 
At node A
 
 
At node
 
 
                
  ?
 
8
 
Assessing
 
Attacker
s problem as seen by the Defender
 
 
 
 
9
 
Assessing
 
10
 
Monte-Carlo approximation of
 
Drawn
 
Generate                    by
 
 
 
 
 
 
 
Approximate
 
11
 
The assessment of
 
The Defender may want to exploit information about
how the Attacker analyzes her problem
 
Hierarchy of decision analysis
Infinite regress
 
12
 
Urban security resource allocation
 
Players
A mayor (Defender)
aims at protecting multiple city districts
The mob (Attacker)
aims at controlling as much value from city districts
Resources to be allocated across the city districts
Mayor
Squad car  
  
 1
Foot patrolmen
 
 2
Mob
Burglars
   
 2
 
 
13
 
Urban security resource allocation
 
City topology and Mayor’s values
 
 
Districts 
 
1
   
2
   
3
Defend-attack-defend setting
Mayor chooses an initial assignment of resources across districts
squad car
patrolmen
The mob sees this and allocates his burglars
Mayor has only time to move the squad car to a neighboring district
at a cost k = 5 units of value
Payoffs from controlling each district are determined at t = 3
Who has control of each district at the end of the game
 
S
i
 = 1  Mayor,  0  Mob;    i = 1,2,3 District
 
t = 1
 
t = 2
 
t = 3
 
A district provides no value
when it is not controlled
 
14
 
Game theoretic approach
 
Common knowledge
Mayor and Mob’s utility and probabilities
Probability that the mayor controls a district
 
15
 
Game theoretic approach
 
Payoffs from controlling districts (additive)
 
Mayor and Mob are risk neutral
 
16
 
Multi-Agent Influence Diagram
 
17
 
Backward induction
 
Mayor best response at t = 3
 
Mob’s best response at t = 2
 
Mayor optimal initial allocation at t = 1
 
18
 
Game Theory solution
 
Optimal decision path
 
City topology and values
 
GT prediction
 
19
 
ARA approach
 
Weaken common knowledge assumption
Mayor does not know with certitude
Mob’s utilities and probabilities
Solve for the mayor (a level-2 thinker)
1.
Assess predictive probability of Mob’s response
                             vs.
2.
Find Mayors allocation (and re-allocation) of MEU
 
20
 
Assessments needed to elicit                 .
 
Mob’s valuations for each city district
                         ,
Prob. of controlling a district
p
D
 = p, but
 
21
 
Assessments needed to elicit                 .
 
Mayor level-2 thinker => Mob level-1 thinker
Mob tries to anticipate mayor’s decision
Mob observes D
1
Mob sees the mayor solving at t = 3
 
Mayor needs to assess
Mob’s estimates of her u
D
 and p
D
 
22
 
Mayor assessments of Mob’s estimates
for her utility and probability
 
Mob’s estimate of her district values v
 
Mob’s estimate of her cost for moving the squad car
 
Mayor beliefs about the mob’s estimate of her utility
 
Mayor beliefs about his estimate of her probabilites
 
23
 
ARA solution
 
MC estimate of
 
Mayor optimal allocation
 
 
 
 
 
 
reallocation response
 
24
 
ARA solution
 
Mob allocates
Squad car to district of most value
(0 0 1)
One patrolman to each of the other two districts
(1 1 0)
She believes that
Other burglar allocations are possible
 
vs. deterministic GT prediction
Mayor moves squad car to (0 1 0) only when
A
2
 = (1 1 0) or (0 2 0)
Which she estimates will happened with prob 0.14
 
GT best response
 
25
 
ARA vs. GT solutions
 
Three best ARA solutions
001|110,  100|011,  010|101
 
cover all districts with some defensive resource
Only one among the 5 best GT solutions does this
Leaving uncover the district of less value for the mayor
GT solution predicts that
the mob will always allocate one of his two burglars
    to the highest value district
 
26
 
Discussion
 
Are available resources to a player known to the
other?
Mob seen as a level-1 thinker who uses GT to find
her allocation decision after estimating her u and p
 Other rationality models for the Mob
Sensitivity analysis
around GT assumption before ARA
focus elicitation around most relevant ARA assessments
 
27
Slide Note
Embed
Share

Adversarial Risk Analysis for Urban Security is a framework aimed at managing risks from the actions of intelligent adversaries in urban security scenarios. The framework employs a Defend-Attack-Defend model where two intelligent players, a Defender and an Attacker, engage in sequential moves, with the Defender deploying defense resources first, and the Attacker choosing an attack in response. By assessing the Adversary's decision problem, probabilities, and utilities, the model aims to predict and respond effectively to potential threats. The approach involves utilizing a SEU model and considering uncertainties in the Attacker's decision-making process.

  • Risk Analysis
  • Urban Security
  • Adversaries
  • Defend-Attack-Defend
  • SEU Model

Uploaded on Sep 06, 2024 | 1 Views


Download Presentation

Please find below an Image/Link to download the presentation.

The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.

E N D

Presentation Transcript


  1. Adversarial Risk Analysis for Urban Security Jesus Rios, IBM Research David Rios, Cesar Gil, D. Banks

  2. Outline Adversarial Risk Analysis The sequential Defend-Attack-Defend Model Urban security resource allocation problem Discussion 2

  3. Adversarial Risk Analysis A framework to manage risks from actions of intelligent adversaries One-sided prescriptive support Use a SEU model Treat the adversary s decision as uncertainties New method to predict adversary s actions We assume the adversary is an expected utility maximizer Model his decision problem Assess his probabilities and utilities Find his action of maximum expected utility But other descriptive models are possible Uncertainty in the Attacker s decision stems from our uncertainty about his probabilities and utilities 3

  4. DefendAttackDefend model Two intelligent players Defender Attacker Sequential moves Defender moves first, deploying defense resources Attacker observes Defender s move and choose an attack Defender moves again responding to the attack 4

  5. DefendAttackDefend model 5

  6. Standard Game Theory Analysis Under common knowledge of utilities and probs At node Expected utilities at node S Best Attacker s decision at node A Best Defender s decision at node Solution: 6

  7. ARA: Supporting the Defender At node A At node ? 8

  8. Assessing Attacker s problem as seen by the Defender 9

  9. Assessing 10

  10. Monte-Carlo approximation of Drawn Generate by Approximate 11

  11. The assessment of The Defender may want to exploit information about how the Attacker analyzes her problem Hierarchy of decision analysis Infinite regress 12

  12. Urban security resource allocation Players A mayor (Defender) aims at protecting multiple city districts The mob (Attacker) aims at controlling as much value from city districts Resources to be allocated across the city districts Mayor Squad car 1 Foot patrolmen 2 Mob Burglars 2 13

  13. Urban security resource allocation City topology and Mayor s values A district provides no value when it is not controlled Districts 1 Defend-attack-defend setting Mayor chooses an initial assignment of resources across districts squad car patrolmen The mob sees this and allocates his burglars Mayor has only time to move the squad car to a neighboring district at a cost k = 5 units of value Payoffs from controlling each district are determined at t = 3 Who has control of each district at the end of the game Si = 1 Mayor, 0 Mob; i = 1,2,3 District 2 3 t = 1 t = 2 t = 3 14

  14. Game theoretic approach Common knowledge Mayor and Mob s utility and probabilities Probability that the mayor controls a district 15

  15. Game theoretic approach Payoffs from controlling districts (additive) Mayor and Mob are risk neutral 16

  16. Multi-Agent Influence Diagram 17

  17. Backward induction Mayor best response at t = 3 Mob s best response at t = 2 Mayor optimal initial allocation at t = 1 18

  18. Game Theory solution GT prediction 1 2 3 City topology and values Optimal decision path 19

  19. ARA approach Weaken common knowledge assumption Mayor does not know with certitude Mob s utilities and probabilities Solve for the mayor (a level-2 thinker) 1. Assess predictive probability of Mob s response vs. 2. Find Mayors allocation (and re-allocation) of MEU 20

  20. Assessments needed to elicit . Mob s valuations for each city district , Prob. of controlling a district pD = p, but 21

  21. Assessments needed to elicit . Mayor level-2 thinker => Mob level-1 thinker Mob tries to anticipate mayor s decision Mob observes D1 Mob sees the mayor solving at t = 3 Mayor needs to assess Mob s estimates of her uD and pD 22

  22. Mayor assessments of Mobs estimates for her utility and probability Mob s estimate of her district values v Mob s estimate of her cost for moving the squad car Mayor beliefs about the mob s estimate of her utility Mayor beliefs about his estimate of her probabilites 23

  23. ARA solution MC estimate of Mayor optimal allocation reallocation response 24

  24. ARA solution Mob allocates Squad car to district of most value (0 0 1) One patrolman to each of the other two districts (1 1 0) She believes that Other burglar allocations are possible vs. deterministic GT prediction Mayor moves squad car to (0 1 0) only when A2 = (1 1 0) or (0 2 0) Which she estimates will happened with prob 0.14 GT best response 25

  25. ARA vs. GT solutions Three best ARA solutions 001|110, 100|011, 010|101 cover all districts with some defensive resource Only one among the 5 best GT solutions does this Leaving uncover the district of less value for the mayor GT solution predicts that the mob will always allocate one of his two burglars to the highest value district 26

  26. Discussion Are available resources to a player known to the other? Mob seen as a level-1 thinker who uses GT to find her allocation decision after estimating her u and p Other rationality models for the Mob Sensitivity analysis around GT assumption before ARA focus elicitation around most relevant ARA assessments 27

Related


More Related Content

giItT1WQy@!-/#giItT1WQy@!-/#giItT1WQy@!-/#