Toward fairness in face matching algorithms

Jamal Alasadi, Ahmed Al Hilli, Vivek K. Singh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Automated face matching algorithms are used in a wide variety of societal applications ranging from access authentication, to criminal identification, to application customization. Hence, it is important for such algorithms to be equitable in their performance for different demographic groups. If the algorithms work well only for certain racial or gender identities, they would adversely affect others. Recent efforts in algorithmic fairness literature (typically not focused on multimedia or computer vision tasks such as face matching) have argued for designing algorithms and architectures to tackle such bias via trade-offs between accuracy and fairness. Here, we show that adopting an adversarial deep learning-based approach allows for the model to maintain the accuracy at face matching while also reducing demographic disparities compared to a baseline (non-adversarial deep learning) approach at face matching. The results motivate and pave way for more accurate and fair face matching algorithms.

Original languageEnglish (US)
Title of host publicationFAT/MM 2019 - Proceedings of the 1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, co-located with MM 2019
PublisherAssociation for Computing Machinery, Inc
Pages19-25
Number of pages7
ISBN (Electronic)9781450369152
DOIs
StatePublished - Oct 15 2019
Event1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, FAT/MM 2019, co-located with ACM Multimedia 2019 - Nice, France
Duration: Oct 25 2019 → …

Publication series

NameFAT/MM 2019 - Proceedings of the 1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, co-located with MM 2019

Conference

Conference1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, FAT/MM 2019, co-located with ACM Multimedia 2019
CountryFrance
CityNice
Period10/25/19 → …

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Hardware and Architecture
  • Signal Processing
  • Software
  • Media Technology

Keywords

  • Algorithmic Fairness
  • Deep Learning.
  • Face Matching

Cite this

Alasadi, J., Al Hilli, A., & Singh, V. K. (2019). Toward fairness in face matching algorithms. In FAT/MM 2019 - Proceedings of the 1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, co-located with MM 2019 (pp. 19-25). (FAT/MM 2019 - Proceedings of the 1st International Workshop on Fairness, Accountability, and Transparency in MultiMedia, co-located with MM 2019). Association for Computing Machinery, Inc. https://doi.org/10.1145/3347447.3356751