|
Bias mitigation for large language models using adversarial learning
Ernst, Jasmina S.
;
Marton, Sascha
;
Brinkmann, Jannik
;
Vellasques, Eduardo
;
Foucard, Damien
;
Kraemer, Martin
;
Lambert, Marian
|
URN:
|
urn:nbn:de:bsz:180-madoc-670249
|
|
Document Type:
|
Conference or workshop publication
|
|
Year of publication:
|
2023
|
|
Book title:
|
Proceedings of the 1st Workshop on Fairness and Bias in AI co-located with 26th European Conference on Artificial Intelligence (ECAI 2023),Kraków, Poland, October 1st, 2023
|
|
The title of a journal, publication series:
|
CEUR Workshop Proceedings
|
|
Volume:
|
3523
|
|
Page range:
|
1-14
|
|
Conference title:
|
1st Workshop on Fairness and Bias in AI
|
|
Location of the conference venue:
|
Kraków, Poland
|
|
Date of the conference:
|
01.10.2023
|
|
Publisher:
|
Calegari, Roberta
;
Tubella, Andrea Aler
;
González Castañe, Gabriel
;
Dignum, Virginia
;
Milano, Michaela
|
|
Place of publication:
|
Aachen, Germany
|
|
Publishing house:
|
RWTH Aachen
|
|
ISSN:
|
1613-0073
|
|
Related URLs:
|
|
|
Publication language:
|
English
|
|
Institution:
|
Außerfakultäre Einrichtungen > Institut für Enterprise Systems (InES)
|
|
Pre-existing license:
|
 Creative Commons Attribution 4.0 International (CC BY 4.0)
|
|
Subject:
|
004 Computer science, internet
|
|
Keywords (English):
|
fairness , debiasing , adversarial learning , NLP , LLMs
|
|
Abstract:
|
Commercial applications increasingly build on large language models (LLMs). Given the inherent biases of LLMs, advancements in fairness research are urgent. Prior methods for mitigating biases in LLMs only address fairness in either language generation tasks or downstream tasks. Additionally, they often incur substantial computational costs by training from scratch. We propose a novel debiasing method that employs adversarial learning during model pre training. Without hyperparameter optimization our comparably computationally efficient method demonstrates increased fairness on a natural language generation task while maintaining performance. In addition, we show that our fairness gains transfer to a downstream task, at a performance cost. We explore a fairness approach which holds a significant potential for redefining the landscape of fairness of LLMs: By learning a single debiased model which can be applied to a variety of tasks, this approach eliminates the need for additional or task-specific debiasing steps. Hence, it facilitates the development of fair commercial applications and constitutes a step towards the broader goal of fairness in societies at large.
|
 | Dieser Eintrag ist Teil der Universitätsbibliographie. |
 | Das Dokument wird vom Publikationsserver der Universitätsbibliothek Mannheim bereitgestellt. |
 Search Authors in
BASE:
Ernst, Jasmina S.
;
Marton, Sascha
;
Brinkmann, Jannik
;
Vellasques, Eduardo
;
Foucard, Damien
;
Kraemer, Martin
;
Lambert, Marian
Google Scholar:
Ernst, Jasmina S.
;
Marton, Sascha
;
Brinkmann, Jannik
;
Vellasques, Eduardo
;
Foucard, Damien
;
Kraemer, Martin
;
Lambert, Marian
ORCID:
Ernst, Jasmina S., Marton, Sascha  ORCID: https://orcid.org/0000-0001-8151-9223, Brinkmann, Jannik, Vellasques, Eduardo, Foucard, Damien, Kraemer, Martin and Lambert, Marian
You have found an error? Please let us know about your desired correction here: E-Mail
Actions (login required)
 |
Show item |
|
![[img]](https://madoc.bib.uni-mannheim.de/style/images/fileicons/application_pdf.png)