Linguistic Models for Analyzing and Detecting Biased Language

By Marta Recasens, Cristian Danescu-Niculescu-Mizil, Dan Jurafsky. 2013

367 citations recorded by the source corpus.

Subfields: Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

Ethical Challenges in Data-Driven Dialogue Systems (2017, background): “…of problems particularly relevant in natural language systems including: predispositions for regional speech patterns (Fine et al. 2014), discriminatory gender bias (Bolukbasi et al. 2016), personal point-of-view bias (Recasens, Danescu-Niculescu-Mizil, and Jurafsky 2013), and numerous others.”
Identifying and Reducing Gender Bias in Word-Level Language Models (2019, background): “Recasens et al. (2013) studies the neutral point of view (NPOV) edit tags in the Wikipedia edit histories to understand linguistic realization of bias.”
UNQOVERing Stereotypical Biases via Underspecified Questions (2020, background): “, 2019, 2020), or an intermediate classification task (Recasens et al., 2013).”
Delete, Retrieve, Generate: a Simple Approach to Sentiment and Style Transfer (2018, background): “While many prior works on linguistic style analysis conﬁrm our observation that attributes often manifest in idiosyncratic phrases (Recasens et al., 2013; Schwartz et al., 2017; Newman et al., 2003), we recognize the fact that in some problems (e.g., Pavlick and Tetreault (2017)), content and…”
Automatically Neutralizing Subjective Bias in Text (2019, methodology): “Workers were shown the Recasens, Danescu-Niculescu-Mizil, and Jurafsky (2013) and Wikipedia deﬁnition of a “biased statement” and six example sentences, then subjected to a ﬁve-question qualiﬁcation test where they had to identify subjectivity bias.”
Political Ideology Detection Using Recursive Neural Networks (2014, background): “More recently, Recasens et al. (2013) detect biased words in sentences using indicator features for bias cues such as hedges and fac-tive verbs in addition to standard bag-of-words and part-of-speech features.”
Predicting Factuality of Reporting and Bias of News Media Sources (2018, methodology): “…pronouns, etc.), and features for clickbait title classiﬁcation from (Chakraborty et al., 2016); • Sentiment : sentiment scores using lexicons (Recasens et al., 2013; Mitchell et al., 2013) and full systems (Hutto and Gilbert, 2014); • Engagement : number of shares, reactions, and comments on…”
A Framework for the Computational Linguistic Analysis of Dehumanization (2020, methodology): “Language is central to studying this phenomenon; like other forms of bias (Wiebe et al., 2004; Greene and Resnik, 2009; Recasens et al., 2013; Voigt et al., 2017; Breitfeller et al., 2019), dehumanizing attitudes are expressed through subtle linguistic manipulations, even in carefullyedited texts.”
Credibility Assessment of Textual Claims on the Web (2016, background): “For example, related work in detecting biased language [17] and credibility analysis in closed communities [12, 11] leverage linguistic features like discourse, subjectivity, and modality.”
Predicting Anti-Asian Hateful Users on Twitter during COVID-19 (2021, methodology): “…following linguistic features by using the News Landscape (NELA) toolkit (Horne et al., 2018): text structure (e.g., POS tags), sentiment, subjectivity, complexity (e.g., readability), bias (Recasens et al., 2013; Mukherjee and Weikum, 2015), and morality (Graham et al., 2009; Lin et al., 2018).”
Language-Aware Truth Assessment of Fact Candidates (2014, background): “Literature on sentiment analysis (Turney et al., 2002; Liu et al., 2005), subjectivity detection (Riloff and Wiebe, 2003; Wiebe et al., 2004), and bias detection (Yu and Hatzivassiloglou, 2003; Recasens et al., 2013) has developed lexicons for identifying subjective language.”
Truth of Varying Shades: Analyzing Language in Fake News and Political Fact-Checking (2017, background): “Hedge words and other vague qualifiers (Choi et al., 2012; Recasens et al., 2013), for example, may add indirectness to a”
Debiasing Word Embeddings from Sentiment Associations in Names (2020, background): “Another line of work aims to detect biased language in text directly using feature-based [12, 26] or neural-based approaches [13, 14].”
Separating Facts from Fiction: Linguistic Models to Classify Suspicious and Trusted News Posts on Twitter (2017, background): “Bias cues Inspired by earlier work on identifying biased language on Wikipedia (Recasens et al., 2013) we extract hedges (expressions of tentativeness and possibility) (Hyland, 2005), assertive verbs (the level of certainty in the complement”
Connotation Frames: A Data-Driven Investigation (2015, background): “There have been growing interests for modeling framing (Greene and Resnik, 2009; Hasan and Ng, 2013), biased language (Recasens et al., 2013) and ideology detection (Yano et al.”
A Transparent Framework for Evaluating Unintended Demographic Bias in Word Embeddings (2019, background): “, 2016, 2017), personal or political view bias (Iyyer et al., 2014; Recasens et al., 2013), and many other examples.”
Bias in machine learning - what is it good for? (2020, background): “In some published work, the word ‘bias’ simply denotes general, usually unwanted, properties of text [25,40].”
A Taxonomy of Knowledge Gaps for Wikimedia Projects (First Draft) (2020, methodology): “Examples include the use of NPOV-templates to automatically detect biased language in the content of Wikipedia [83, 137], or documenting biases in specific topics such as politics [73] or culture [20] among others.”
Stance and Sentiment in Tweets (2016, background): “me of elections [Tumasjan et al. 2010; Bermingham and Smeaton 2011; Lampos et al. 2013]. There are other subtasks in opinion mining related to stance classiﬁcation, such as biased language detection [Recasens et al. 2013; Yano et al. 2010], perspective identiﬁcation [Lin et al. 2006] and user classi”