4min chapter

Responsibly Deploy AI in Production with Anupam Datta

Machine Learning Archives - Software Engineering Daily

CHAPTER

Is There a Disparity in the Performance of Machine Learning Models?

There is a potential here for large scales asital harm. Machine learning models will pick op and reinforce historical bias in training data, the data on which they are being trained. One way to address this problem is to ensure that the training data is diverse and representative so that accuracy levels are balanced across different groups. There can be other kinds of rood causes. Let's think about an example where the model is excepting men at a higher rate than women. This is one crete example of where the root causeis and how to address it.

00:00