Machine Learning and Fairness
Written by Marwa Khalil
This month we were back with the first event since our summer break. We were back at the Google office to hear about machine learning and fairness.
Our speaker Silvia Chiappa, senior research scientist at DeepMind, gave us a crash course on machine learning and guided us through the ways in which it has been used in the past. Silvia spoke about how and why machine learning has gone wrong and spoke on the fairness of machine learning.
Machine learning is an approach to AI where the machine system is given a mathematical framework to extract the rules from data. A machine learning algorithm is unfair if it produces unjust or prejudice results. This commonly occurs with regards to sensitive attributes such as race and gender.
Machine learning is essentially only as good as the data that it is based on and problems arise when humans lose control. Control can be lost when the data is heavily biased, this means that the expected predications can be wrong or worse, harmful to society. This creates unfairness in the systems that rely on machine learning.
Even if biases are removed from the data and algorithms there are other problems:
• More and better quality data is required to build fairer models
• Fairness is objective
• Fairness is not a quantitive concept
• The machine learning systems are problem specific
• Machines do not follow human legislation
• Training may be required to understand the models
After Silvia’s presentation there was a workshop and discussion based on questions that were raised by Silvia. This sparked lively discussions within the groups and highlighted the different perspectives. The presentation and discussions were thought provoking and introduced an important concept in UX of the future.
Speaker
Silvia Chiappa is a senior research scientist at DeepMind, where she works on deep predictive models of high-dimensional time-series, and also contributes to the DeepMind’s diversity and inclusion initiative. Silvia received a Diploma di Laurea in Mathematics from University of Bologna and a PhD in Statistical Machine Learning from École Polytechnique Fédérale de Lausanne. Before joining DeepMind, she worked in several Machine Learning and Statistics research groups, such as the Empirical Inference Group at the Max-Planck Institute for Biological Cybernetics, the Machine Learning and Perception Group at Microsoft Research Cambridge, and the Statistical Laboratory at University of Cambridge. Silvia’s research interests are based around Bayesian and causal reasoning, approximate inference, time-series models, algorithmic fairness, and deep learning.
Thanks
A huge thank you to Google for hosting us and the speaker for the engaging talk and discussion.
See more photos of the event on our Flickr album
Contributing to LTUX London
Do you fancy writing for Ladies that UX London? If so we would love to hear your ideas.
We are always looking for creative, interesting and informative contributions to the Ladies that UX London blog and community. Wether you want to inspire other ladies in the UX industry or share your expertise; we’re all ears.
Never miss a trick
🎉 Never miss a trick. Get the latest updates, priority tickets and exclusive offers by signing up to our newsletter.
☎️ Want to get in contact with us? You can find us on Twitter, Facebook and Instagram.
💖 If you like this post share the love and give us a round of applause.
😻 If you have anything you want featured, why not get in touch? We would love to hear from you.
About the Author
Marwa Khalil is a London based UX and UI designer who has had the chance to work with some great London startups and SaaS companies. She is a regular contributor and editor for the LTUX London blog. Stay up to date with her on Medium, Twitter and Instagram.
