Breakthrough Paper on Fair AI Wins Top Award at ACM FAccT 2023

What is it about?

Chicago, June 12, 2023 – A pioneering paper by Zichong Wang and his team, under the guidance of advisor Wenbin Zhang, has won the Best Paper Award at the ACM Conference on Fairness, Accountability, and Transparency (FAccT) 2023. The paper, titled "Preventing Discriminatory Decision-making in Evolving Data Streams," introduces a new method to ensure fairness in machine learning, especially when dealing with constantly changing data.

Featured Image

Photo by Tingey Injury Law Firm on Unsplash

Photo by Tingey Injury Law Firm on Unsplash

Why is it important?

The team's method, called Fair Sampling over Stream (FS2), is designed to keep machine learning models fair and accurate even as new data comes in continuously. They also developed a new metric called Fair Bonded Utility (FBU) to measure how well the method balances fairness and performance. "Real-world applications often involve data that changes over time, and our method addresses these changes to keep models fair," said Wang during his presentation. Their research shows that FS2 performs better than existing methods in maintaining fairness without sacrificing accuracy. This is crucial as more industries rely on machine learning for decisions that can impact people's lives. The ACM FAccT 2023 conference in Chicago gathered experts to discuss the ethical use of AI and machine learning. Wang's award-winning paper highlights the ongoing effort to create fair and accountable AI systems. For more details on the research, you can read the full paper in the ACM Digital Library below.

Perspectives