Grad Student Zichong Wang Receives Best Paper Award at ACM FAccT Conference

Zichong Wang at 2023 FAccT Conference with Best Paper Award
Zichong Wang (left) with Best Paper Award from 2023 FAccT Conference
Zichong Wang presents his paper at 2023 FAccT conference

A paper co-authored by graduate student Zichong Wang (Computer Science) was awarded the Best Paper Award at the sixth annual ACM FAccT Conference, which took place in Chicago, Illinois, June 12-15, 2023. Read and download the paper here.

Zichong Wang is advised by Assistant Professor Wenbin Zhang (Computer Science).

The title of the paper, presented in the FAccT Conference Session 14 (Practical Methods),” is “Preventing Discriminatory Decision-making in Evolving Data Streams.” Read the paper’s abstract below.

Wang’s Michigan Tech co-authors are faculty members Wenbin Zhang (Computer Science), Dukka KC (Computer Science), and Shan Zhou (Social Sciences), and students Tyler Zetty (Computer Engineering) and Sneha Karki (Data Science).

Additional co-authors are Nripsuta Saxena, University of Southern California; Tongjia Yu, Columbia University; Israat Haquem, Dalhousie University, Canada; Ian Stockwell, University of Maryland; Xuyu Wang, Florida International University; and Albert Bifet, University of Waikato, New Zealand.

Talk Abstract: Bias in machine learning has rightly received significant attention over the past decade. However, most fair machine learning (fair-ML) works to address bias in decision-making systems has focused solely on the offline setting. Despite the wide prevalence of online systems in the real world, work on identifying and correcting bias in the online setting is severely lacking. The unique challenges of the online environment make addressing bias more difficult than in the offline setting. First, Streaming Machine Learning (SML) algorithms must deal with the constantly evolving real-time data stream. Secondly, they need to adapt to changing data distributions (concept drift) to make accurate predictions on new incoming data. Incorporating fairness constraints into this already intricate task is not straightforward. In this work, we focus on the challenges of achieving fairness in biased data streams while accounting for the presence of concept drift, accessing one sample at a time. We present Fair Sampling over Stream (FS2), a novel fair rebalancing approach capable of being integrated with SML classification algorithms. Furthermore, we devise the first unified performance-fairness metric, Fairness Bonded Utility (FBU), to efficiently evaluate and compare the trade-offs between performance and fairness across various bias mitigation methods. FBU simplifies the comparison of fairness-performance trade-offs of multiple techniques through one unified and intuitive evaluation, allowing model designers to easily choose a technique. Overall, extensive evaluations show our measures surpass those of other fair online techniques previously reported in the literature.

The annual ACM FAccT Conference brings together researchers and practitioners interested in fairness, accountability, and transparency in socio-technical systems.