Write a Blog >>
ICSE 2021
Mon 17 May - Sat 5 June 2021

Machine learning software can be unfair when making human-related decisions, having prejudices over certain groups of people. Existing work primarily focuses on proposing fairness metrics and presenting fairness improvement approaches. It remains unclear how key aspect of any machine learning system, such as the feature set and training data, affect fairness. This paper presents results from a comprehensive study that addresses this problem. We find that extending the feature set plays a significant role in improving fairness (with an average increase rate of 38%). Importantly, and contrary to widely-held beliefs that greater fairness often corresponds to lower accuracy, our findings reveal that extending the feature set improves both accuracy and fairness. Perhaps also surprisingly, we find that extending training data does not improve fairness. Our results suggest such data enhancement can further reduce fairness when feature sets are insufficient; an important cautionary finding for practising software engineers.

Conference Day
Thu 27 May

Displayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change

10:00 - 10:55
3.1.1. Social Equality and Fairness #2Journal-First Papers / SEIS - Software Engineering in Society / Technical Track at Blended Sessions Room 1 +12h
Chair(s): Alexander SerebrenikEindhoven University of Technology
10:00
15m
Paper
A First Look at Human Values-Violation in App ReviewsSEIS
SEIS - Software Engineering in Society
Humphrey ObieMonash University, Waqar HussainMonash University, Xin XiaHuawei Software Engineering Application Technology Lab, John GrundyMonash University, Li LiMonash University, Burak TurhanMonash University, Jon WhittleCSIRO's Data61 and Monash University, Mojtaba ShahinMonash University
Pre-print
10:15
20m
Full-paper
"Ignorance and Prejudice" in Software FairnessTechnical Track
Technical Track
Jie M. ZhangUCL, Mark HarmanUniversity College London
Pre-print
10:35
20m
Paper
Gender Differences in Personality Traits of Software EngineersJournal-First
Journal-First Papers
Daniel RussoDepartment of Computer Science, Aalborg University, Klaas-Jan StolUniversity College Cork and Lero, Ireland
Link to publication DOI Pre-print
22:00 - 22:55
22:00
15m
Paper
A First Look at Human Values-Violation in App ReviewsSEIS
SEIS - Software Engineering in Society
Humphrey ObieMonash University, Waqar HussainMonash University, Xin XiaHuawei Software Engineering Application Technology Lab, John GrundyMonash University, Li LiMonash University, Burak TurhanMonash University, Jon WhittleCSIRO's Data61 and Monash University, Mojtaba ShahinMonash University
Pre-print
22:15
20m
Full-paper
"Ignorance and Prejudice" in Software FairnessTechnical Track
Technical Track
Jie M. ZhangUCL, Mark HarmanUniversity College London
Pre-print
22:35
20m
Paper
Gender Differences in Personality Traits of Software EngineersJournal-First
Journal-First Papers
Daniel RussoDepartment of Computer Science, Aalborg University, Klaas-Jan StolUniversity College Cork and Lero, Ireland
Link to publication DOI Pre-print