logseq/pages/@Towards model-based bias mitigation in machine learning.md

tags:: readingnotes date:: 23-10-2022 publisher:: ACM place:: Montreal Quebec Canada conference-name:: MODELS '22: ACM/IEEE 25th International Conference on Model Driven Engineering Languages and Systems proceedings-title:: Proceedings of the 25th International Conference on Model Driven Engineering Languages and Systems isbn:: 978-1-4503-9466-6 doi:: 10.1145/3550355.3552401 title:: @Towards model-based bias mitigation in machine learning pages:: 143-153 item-type:: ConferencePaper access-date:: 2023-04-20T13:13:07Z original-title:: Towards model-based bias mitigation in machine learning language:: en url:: https://dl.acm.org/doi/10.1145/3550355.3552401 authors:: Alfa Yohannis, Dimitris Kolovos library-catalog:: DOI.org (Crossref) links:: Local library, Web library

• Abstract
  • Models produced by machine learning are not guaranteed to be free from bias, particularly when trained and tested with data produced in discriminatory environments. The bias can be unethical, mainly when the data contains sensitive attributes, such as sex, race, age, etc. Some approaches have contributed to mitigating such biases by providing bias metrics and mitigation algorithms. The challenge is users have to implement their code in general/statistical programming languages, which can be demanding for users with little programming and fairness in machine learning experience. We present FairML, a model-based approach to facilitate bias measurement and mitigation with reduced software development effort. Our evaluation shows that FairML requires fewer lines of code to produce comparable measurement values to the ones produced by the baseline code.
• Attachments
  • Yohannis e Kolovos - 2022 - Towards model-based bias mitigation in machine lea.pdf {{zotero-imported-file YKTMX6QH, "Yohannis e Kolovos - 2022 - Towards model-based bias mitigation in machine lea.pdf"}}
• Highlights
  • ((6443aff4-d368-41c5-8e3f-0fab050a9696))
  • ((6443b029-7c3c-4976-bdb2-636d041bfb2f))
  • ((6443b0c6-c048-4241-a9e7-a72164d9abf4))
    • ((644a3766-d009-4482-a078-0d804d5c6544))
  • ((644a3855-1187-418f-bf42-23cb6705a978))
    • question ((644a3871-b847-40e7-ab54-d14701ae1758))
      • What's the different with the proposed fairness assessment approach in PAPERS/FAIRNESS ?
  • ((644a3b6a-35ac-4991-8cc8-ef1568904f34))
  • ((644a3bf7-d6c6-496d-af65-4a854829aa7d))
    • Se the proposed approach is for automatically mitigating bias by putting in place different metrics for particular situations
      • ((644a3c42-6144-4d05-9317-9f85ea65bca7))
    • ((644a3c81-dd2f-4e74-a637-d0cb20e328bb))
      • question For instance these are metrics to measure equal fairness. Isn't this related to our fariness assessment approach presented in PAPERS/FAIRNESS?
  • ((644a3e1b-78e8-4449-a717-d97ba70e719f))
  • ((644a3f06-4ed0-4d0b-ac12-d922c6be39e9))
    • ((644a3f9a-2318-4e82-9f20-dca4539de90b))
      • question how are we different in PAPERS/FAIRNESS with all of these?