Identifying discriminatory patterns in online ad data
We're teaming up with the ARC Centre of Excellence for Automated Decision Making & Society to bring you the winners from the recent ADM+S Dark Ads Hackathon.
The winning team used three large datasets to identify discriminatory patterns such as price discrimination or proxy discrimination — where targeting based on one attribute (like geographic location) creates discrimination based on another such as income, race or cultural background.
The datasets included the ADM+S Australian Ad Observatory dataset (500,000+ ads donated by 2,000 Australian Facebook users), Facebook’s WAIST (why am I seeing this?) data, and statistical data associated with postcodes.
The team will discuss the tools they developed, patterns already found, and plans for future work and development of their idea.
This discussion will be moderated by co-lead of the ADM+S Dark Ads project and hackathon, and Hacks/Hackers Brisbane member, Professor Daniel Angus.
When: Friday, November 25
Time: 5pm for 5:30pm start
Where: Foyer of the ABC building, South Bank
🌳 A Visual Bibliography of Tree Visualisation 2.0: The endless variety of tree data visualisations is, frankly, astonishing. This is a collection of more than 300 different techniques sourced mostly from academic literature.
🤖 Notes on writing with Google’s AI : Robin Sloan’s newsletter is well worth a subscribe, but the latest one really jumped out as relevant to this newsletter’s audience. If you’ve got any interest in the intersection of art, writing and machine learning, check it out.
Big thanks to Simon Elvery for speaking to us in September about his work diving into the modern reporter’s notebook.
Earlier this year, Simon spent three months at the Reuters Institute for the Study of Journalism in the UK, where he reviewed literature and conducted interviews with 12 reporters from seven countries about the tools they use in their practice and what stands in the way of incorporating new tools into their standard kit.
He found that reporters are not eager early adopters of new tech and often make do with what they're familiar with or have easy access to - and usually that means settling for something that is not made with journalism in mind.
Thanks for reading Hacks/Hackers Brisbane! Subscribe for free to receive new posts and support my work.