Game Theory For Computer Science

2481 Words10 Pages
Temp - A Game Theoretical Model for Adversarial Learning
Duregsh Pandey


What is the summary of the proposed scheme? 3
How game theory is applied in this work? 4
Your thoughts on other aspects of Game theory that can be applied for this problem 5
Your assessment about the paper 6

1. What is the summary of the proposed scheme? (5 marks)
This scheme first, introduces that how adversaries are able to affect the accuracy of the data miner by manipulating their data, and then it models the interaction between the adversary and the data miner as a two-person sequential non cooperative Stackelberg game and analyze the outcomes when there is a natural leader and a follower. It then proceed
…show more content…
There is a very old model of the classifiers and adversaries, where it is assumed that both of them have full information about data, and the classifiers set some boundaries on the data D and attacker recognizes the those boundaries and modifies the data to D’, again classifiers set new boundaries, and in this manner, the process goes till infinite time and thus a sequential game between the classifier and the attacker has been establishes. However, the key idea in game theory is that of equilibrium: a state from which neither the classifier nor the adversary will have any incentive to deviate.
The scheme represents the model of the “spammer and data miner game”, in which the strategy of the spammer and data miner has been explained. After this, the scheme also explains what may be the possible outcomes of this model and whether we can get the equilibrium state in this model if we are using the backward induction.
The Model of “Infinite stackelberg game” has been introduced and explained in this paper. The rollback or the backward induction can be seen as an optimization problem. The classification problem can be modeled as a game theoretic problem.
Genetic algorithm for solving the Infinite stackelberg game has been explained. The gist of the
Get Access