On the Unintended Social Bias of Training Language Generation Models with Data from Local Media
By Omar U. Florez. 2019
5 citations recorded by the source corpus.
There are concerns that neural language models may preserve some of the stereotypes of the underlying societies that generate the large corpora needed to train these models. For example, gender bias is a significant problem when generating text, and its unintended memorization could impact the user experience of many applications (e.g., the smart-compose feature in Gmail). In this paper, we introduce a novel architecture that decouples the representation learning of a neural model from its memory management role. This architecture allows us to update a memory module with an equal ratio across gender types addressing biased correlations directly in the latent space. We experimentally show that our approach can mitigate the gender bias amplification in the automatic generation of articles news while providing similar perplexity values when extending the Sequence2Sequence architecture.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
In Computer Vision and Pattern Recognition
Citations in the corpus
- Language (Technology) is Power: A Critical Survey of “Bias” in NLP (2020, neutral): “Vague/unstated Rudinger et al. (2018); Webster et al. (2018); Dinan et al. (2019); Florez (2019); Jumelet et al. (2019); Lauscher et al. (2019); Liang et al. (2019); Maudslay et al. (2019); May et al. (2019); Prates et al. (2019); Prost et al. (2019); Qian et al. (2019); Swinger et al. (2019); Zhao…”
- What’s in the Box? An Analysis of Undesirable Content in the Common Crawl Corpus (2021, background): “However, these increasingly complex models also require increasingly large amounts of data to train them, which is often a combination of curated, high-quality datasets such as encyclopedic articles and books and non-curated content from the Web (Radford et al., 2018, 2019).”