Bias in decisions made by modern software is becoming a common and serious problem. We present Themis, an automated test suite generator to measure two types of discrimination, including causal relationships between sensitive inputs and program behavior. We explain how Themis can measure discrimination and aid its debugging, describe a set of optimizations Themis uses to reduce test suite size, and demonstrate Themis’ effectiveness on open-source software. Themis is open-source and all our evaluation data are available at http://fairness.cs.umass.edu/. See a video of Themis in action: https://youtu.be/brB8wkaUesY
Tue 6 NovDisplayed time zone: Guadalajara, Mexico City, Monterrey change
Tue 6 Nov
Displayed time zone: Guadalajara, Mexico City, Monterrey change
10:30 - 12:00 | |||
10:30 22mTalk | BigSift: Automated Debugging of Big Data Analytics in Data-Intensive Scalable Computing Demonstrations Muhammad Ali Gulzar University of California, Los Angeles, Siman Wang Hunan University, Miryung Kim University of California, Los Angeles | ||
10:52 22mTalk | FOT: A Versatile, Configurable, Extensible Fuzzing Framework Demonstrations Hongxu Chen Nanyang Technological University, Yuekang Li Nanyang Technological University, Bihuan Chen Fudan University, Yinxing Xue , Yang Liu Nanyang Technological University, Singapore | ||
11:15 22mTalk | Themis: Automatically Testing Software for Discrimination Demonstrations Rico Angell University of Massachusetts, Amherst, Brittany Johnson University of Massachusetts Amherst, Yuriy Brun University of Massachusetts Amherst, Alexandra Meliou University of Massachusetts Amherst Link to publication DOI Pre-print | ||
11:37 22mDemonstration | Vista: Web Test Repair Using Computer Vision Demonstrations Andrea Stocco University of British Columbia, Rahulkrishna Yandrapally University of British Columbia, Canada, Ali Mesbah University of British Columbia Pre-print | ||