Today, in our final lab, Jabril tries to make an AI to settle the question once and for all, "Will a cat or a dog make us happier?" But in building this AI, Jabril will accidentally incorporate the very bias he was trying to avoid. So...

MIT grad student Joy Buolamwini was working with facial analysis software when she noticed a problem: the software didn't detect her face -- because the people who coded the algorithm hadn't taught it to identify a broad range of skin...

Today, we're going to talk about five common types of algorithmic bias we should pay attention to: data that reflects existing biases, unbalanced classes in training data, data that doesn't capture the right value, data that is amplified...

Pupil outcome: I can describe algorithmic bias and suggest ways to make algorithms fairer. Key learning points: - Algorithmic bias is when an algorithm produces unfair or discriminatory outcomes that favour some groups over others. -...

In the second part of this series on Algorithmic Bias and Fairness, we're looking at how we can make artificial intelligence and algorithms fairer.

Does Algorithmic Fairness Include the LGBTQ+ Community?

Is Sara Hooker concerned about the short term implications of ML models, AI in general or the risks of AGI (Artificial General Intelligence) and the skynet idea and it having drastic consequences for human survival?

Machine Learning...

Learn of fundamental architectural constraints and the patterns that most models find important and how fundamentally different they are from the patterns a human would expect to be important.

Machine Learning Systems Design with...

Does Sara Hooker think effective tools should look primarily at the data or the model itself?<br/>
Machine Learning Systems Design with Sara Hooker, Part 6

Learn how Sara Hooker balances pragmatic considerations and understanding the algorithms we use. How has she seen interpretability methods adapted and what work has she done with interpretability?

Machine Learning Systems Design...

Sara Hooker elaborates on the methods applied to saliency, interpretability methods in order to find the flaws in them.<br/>
Machine Learning Systems Design with Sara Hooker, Part 2

Learn how we should look less at the data and make changes to the model themselves on a per bias basis.<br/>
Machine Learning Systems Design with Sara Hooker, Part 5

Are there more semantic reasons behind targeting compactness? Does Sara Hooker believe that deep neural networks are a solution in the long term?<br/>
Machine Learning Systems Design with Sara Hooker, Part 7

Sara Hooker explains why algorithmic bias is not a data collection problem and what are the implicit biases of the algorithms themselves.<br/>
Machine Learning Systems Design with Sara Hooker, Part 4

Sara Hooker gives her opinion about other techniques like probabilistic models or explicit knowledge based models, and whether she thinks there will be a resurgence and if it will be valuable in the long term.

Machine Learning...