Twitter is partnering with UC Berkeley researchers in an effort to improve its machine learning capabilities.
Twitter uses machine learning to onboard users, prepare timelines, and manage the overall ad and user experience.
Sandeep Pandey, senior director of software engineering at Twitter said in a blog post that “making Twitter more healthy requires making the way we practice [machine learning] more fair, accountable and transparent.”
The newly announced research initiative will focus on improving the performance of machine learning across social networks such as Twitter and is led by Professor Moritz Hardt and Professor Ben Recht.
UC Berkeley will closely collaborate with the social network to gain a greater understanding of how machine learning algorithms impact people. This includes behavioural responses to recommendations, which in turn leads to a shift in what the algorithm learns.
Twitter has previously sponsored initiative such as the ACM FAT* 2019 conference which focuses on the societal impact of machine learning.