Visual Exploration of Indirect Biases in Natural Language Processing Transformer Models | TU Wien

Information

Publication Type: Master Thesis
Workgroup(s)/Project(s): not specified
Date: 2022
Date (Start): January 2022
Date (End): November 2022
Diploma Examination: 21. November 2022
Open Access: yes
First Supervisor: Manuela Waldner
Pages: 134
Keywords: visual exploration, visual analytics, natural language processing, bias, transformer models

Abstract

In recent years, the importance of Natural Language Processing has been increasing with more and more fields of application. The word representations, such as word embedding or transformer models, used to transcribe the language are trained using large text corpora that may include stereotypes. These stereotypes may be learned by Natural Language Processing algorithms and lead to biases in their results. Extensive research has been performed on the detection, repair and visualization of the biases in the field of Natural Language Processing. Nevertheless, the methods developed so far mostly focus on word embeddings, or direct and binary biases.To fill the research gap regarding multi-class indirect biases learned by transformer models, this thesis proposes new visualisation interfaces to explore indirect and multi-class biases learned by BERT and XLNet models. These visualisations are based on an indirect quantitative method to measure the potential biases encapsulated in transformer models, the Indirect Logarithmic Probability Bias Score. This metric is adapted from an existing one, to enable the investigation of indirect biases. The evaluation of our new indirect method shows that it enables to reveal known biases and to discover new insights which could not be found using the direct method. Moreover, the user study performed on our visualization interfaces demonstrates that the visualizations supports the exploration of multi-class indirect biases, even though improvements may be needed to fully assist the investigation of the sources of the biases.

Additional Files and Images

Additional images and videos

teaser

Additional files

Master thesis

Weblinks

BibTeX

@mastersthesis{louis-alexandre_dit_petit-frere-2022-veo,
  title =      "Visual Exploration of Indirect Biases in Natural Language
               Processing Transformer Models",
  author =     "Judith Louis-Alexandre Dit Petit-Frere",
  year =       "2022",
  abstract =   "In recent years, the importance of Natural Language
               Processing has been increasing with more and more fields of
               application. The word representations, such as word
               embedding or transformer models, used to transcribe the
               language are trained using large text corpora that may
               include stereotypes. These stereotypes may be learned by
               Natural Language Processing algorithms and lead to biases in
               their results. Extensive research has been performed on the
               detection, repair and visualization of the biases in the
               field of Natural Language Processing. Nevertheless, the
               methods developed so far mostly focus on word embeddings, or
               direct and binary biases.To fill the research gap regarding
               multi-class indirect biases learned by transformer models,
               this thesis proposes new visualisation interfaces to explore
               indirect and multi-class biases learned by BERT and XLNet
               models. These visualisations are based on an indirect
               quantitative method to measure the potential biases
               encapsulated in transformer models, the Indirect Logarithmic
               Probability Bias Score. This metric is adapted from an
               existing one, to enable the investigation of indirect
               biases. The evaluation of our new indirect method shows that
               it enables to reveal known biases and to discover new
               insights which could not be found using the direct method.
               Moreover, the user study performed on our visualization
               interfaces demonstrates that the visualizations supports the
               exploration of multi-class indirect biases, even though
               improvements may be needed to fully assist the investigation
               of the sources of the biases.",
  pages =      "134",
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien",
  keywords =   "visual exploration, visual analytics, natural language
               processing, bias, transformer models",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2022/louis-alexandre_dit_petit-frere-2022-veo/",
}