Last Week in AI

Every week, my team at Invector Labs publishes a newsletter to track the most recent developments in AI research and technology. You can find this week’s issue below. You can sign up for it below. Please do so, our guys worked really hard on this:

From the Editor: Fairness in AI

One of the greatest benefits of artificial intelligence(AI) is that it can produce results that are pragmatic and not vulnerable to the subjectivity and biases of humans. However, that statement of only partially true. While AI systems don’t make decisions based on feelings or emotions, they do inherit a lot of human biases via the training datasets. Bias is relevant because it leads to unfairness. Defining a notion of fairness is one of the most important challenges of the next decade of AI.

Building more fairer AI systems seems like an obvious goal but how to we exactly define fairness. Conceptually, fairness can be define as a relationship between a sensitive input attribute such as race or gender and the output of the model. However, that’s just part of the equation. AI Fairness is directly related to the biases in the data generation model. This week, DeepMind published some clever work about how to build more fairer machine learning systems using a very old statistical model. DeepMind’s work is an example of the importance of the role that fairness will play in…