Debiasing

Mitigating Gender Bias in Natural Language Processing: Literature Review

by

Author: Tony Sun, Andrew Gaut, Shirlyn Tang, et al.; Publisher: Arxiv; Publication Year: N/A. The following article discusses how despite their success in modeling various applications, natural language processing (NLP) models propagate and may even amplify gender bias found in text corpora. While the study of bias in artificial intelligence is not new, methods to reduce gender bias in NLP are still in their early stages. The authors of this paper…

Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings

by

Author: Tolga Bolukbasi, Kai-Wei Chang, James Zou, et al.; Publisher: Arxiv; Publication Year: 2016. The following article discusses how machine learning applied blindly runs the risk of amplifying biases in data. Word embedding, a popular framework for representing text data as vectors that has been used in many machine learning and natural language processing tasks, poses such a risk. The authors demonstrate that even word embeddings…

Using Unethical Data to Build a More Ethical World

by

Author: Jaime Brandon; Publisher: Springer Link; Publication Year: 2020. The following article discusses how it is commonly known that there are lots of algorithms that are biased in one way or another. In this paper, the researchers did not block any bias from forming in their model to see what they can learn from that process. The article states that transparency with biases instigates change. By choosing not to debias the…

Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them

by

Author: Hila Gonen, Yoav Goldberg; Publisher: Allen Institute for Artificial Intelligence; Publication Year: 2019. The following study talks about gender bias in word embeddings, an important component in Natural Language Processing. There have been debiasing methods that were created and widely used to help eliminate the bias, but the paper argues that current debiasing methods are superficial and are just hiding the bias, not completely eliminating it…

Bias in AI: What it is, Types, Examples & 6 Ways to Fix It in 2022

by

Author: Cem Dilmegani; Publisher: AI Multiple; Publication Year: 2022. The following article describes how in the imperfect world, AI can’t be expected to be completely unbiased. However, there are various ways to minimize bias by proper testing of algorithms and building AI systems with responsible AI principles. Fixing bias in AI systems can be summarized in the following 6 steps: 1)…

Detecting and Mitigating Bias in Natural Language Processing

by

Author: Aylin Caliskan; Publisher: Brookings; Publication Year: 2021. The following article looks at how billions of people using the internet every day are exposed to biased word embeddings. However, word embedding debiasing is not a feasible solution to the bias problems since debiasing word embeddings remove essential context about the world. Instead of blindly debiasing…