Author: Jonathan Keith Lee Jarvi

Machine learning is being used in countless different fields for a variety of different purposes. In order to develop the necessary machine learning models for these applications researches need to gather large amounts of data to train and test the model on. As with any other field, there will be malicious attackers who wish to compromise the model. The attacks where the malicious attacker tries to control a portion of the data used to train the model is called a poisoning attack. These malicious data points may hinder the accuracy of the model and prevent the model from properly analyzing the genuine data.

Some researchers have decided to use game theory to model the poisoning attacks and they were able to conclude that there is no pure strategy Nash equilibrium for this problem. As seen in the figure above after around 20% of the data is removed the defense does not improve the accuracy of the model, and instead harms the accuracy as the amount removed increases. Instead they found a mixed strategy Nash equilibrium for the attacker which they better protected the model and maintained a higher accuracy.

Resources:

https://techxplore.com/news/2019-06-game-theory-poisoning-scenarios.html

Y. Ou and R. Samavi, “Mixed Strategy Game Model Against Data Poisoning Attacks,” 2019 49th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W), Portland, OR, USA, 2019, pp. 39-43.

Countless device in our dailies lives use electricity. This number continues to increase as time passes by. To power these new devices we need ever more energy. The main source of electricity for the majority of devices is from a power grid. These massive physical networks can span across different countries and provide electricity for the majority of the world.

For a graph of a power grid we can take the … to be nodes and the cables between them to be edges. We can also measure the electrical current through an edge as the weight of that edge. We can use graph flow methods to measure the importance of different edges, and asses whether the given infrastructure is enough to handle the load. Occasionally there can be blackouts, which can occur as the result of some edge between two nodes being removed. Larger blackout can occur as the result of cascading failures occurring after several edges fail. These failures can result in increased flow on several edges, which may lead to these edges failing as well increasing the size of the blackout.

Sources:

What was behind Friday’s National Grid outage? Network theory, not conspiracy