LTUX - Machine Learning and Fairness
This weeks LTUX event, Machine Learning and Fairness, was held at the Google offices and with a title like that I knew we would be in for a challenging evening; I was not wrong.
Silvia Chiappa was our speaker and started by stating that she does not have a recipe for how to make machine learning fair, but wanted to make us more aware of the issues so we can think about it when designing.
Do you know what machine learning is? "Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed."
Can you think of anything in your life that uses machine learning? What about voice assistants like Siri; most web searches; detection of credit card fraud; those ads that follow you round the web. There are many more examples out there.
There are so many problems with machine learning. For instance many people think machine learning is perfect, however it is only as good as the data we give it and the algorithms we help it create, and as all humans are biased and flawed, even when we try our best, this does not bode well for machine learning.
For instance there is a facial recognition system that does not recognise black faces because it has been built and tested on white people.
Another problem seems to be around regulations for this technology, the data it has access to and how rigorously it is tested.
So if you then ask if it is fair you can guess the answer.
Add to this the fact we do not always understand how it came up with the result it did and we find it much harder to correct the issues.
Having listened to Silvia's talk we then split into groups to discuss what machine learning we have come across in our day-to-day lifes or in our jobs. This generated a lot of discussion and I think we all started to realise that there are not many parts of our lives where machine learning is not touching somewhere.
This was a great evening and I liked that we did not just get to hear Silvia speak but also got to discuss this massive, complicated topic too. A great way to aid our understanding!