Empath: Understanding Topic Signals in Large-Scale Text

By Ethan Fast, Binbin Chen, Michael S. Bernstein. 2016

441 citations recorded by the source corpus.

Human language is colored by a broad range of topics, but existing text analysis tools only focus on a small number of them. We present Empath, a tool that can generate and validate new lexical categories on demand from a small set of seed terms (like "bleed" and "punch" to generate the category violence). Empath draws connotations between words and phrases by deep learning a neural embedding across more than 1.8 billion words of modern fiction. Given a small set of seed words that characterize a category, Empath uses its neural embedding to discover new related terms, then validates the category with a crowd-powered filter. Empath also analyzes text across 200 built-in, pre-validated categories we have generated from common topics in our web dataset, like neglect, government, and social media. We show that Empath's data-driven, human validated categories are highly correlated (r=0.906) with similar categories in LIWC.

Subfields: Artificial Intelligence

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

Citations in the corpus

Lawyers are Dishonest? Quantifying Representational Harms in Commonsense Knowledge Resources (2021, methodology): “collect a list of biased words that could represent favoritism and prejudice from LIWC (Tausczik and Pennebaker, 2010) and Empath (Fast et al., 2016).”
Long-Term Trends in the Public Perception of Artificial Intelligence (2016, background): “For example, it is possible to crowdsource lexicons of words to answer novel research questions (Fast, Chen, and Bernstein 2016), or leverage crowds to bootstrap classiﬁers that can then be applied to much larger corpora (Danescu-Niculescu-Mizil et al. 2013; Fast and Horvitz 2016).”
The Truth is Out There: Investigating Conspiracy Theories in Text Generation (2021, methodology): “We utilize Empath (Fast et al., 2016 lexical categories to retrieve the top linguistic topics across the three GPT-2 text sets and Reddit dataset.”
Something’s Brewing! Early Prediction of Controversy-causing Posts from Discussion Features (2019, background): “…links, number of imgur links, number of sentences, Flesch-Kincaid readability score, rate of italics, rate of bold-face, presence of a list, and the rate of word use from 25 Empath wordlists (Fast et al., 2016), which include various categories, such as politeness, swearing, sadness, etc. W2V .”
Quantifying Engagement with Citations on Wikipedia (2020, methodology): “Figure 12: Empath [14] topics most strongly (anti-)associated with citation events (cf.”
A survey of multimodal sentiment analysis (2017, background): “While the inference of word categories based on deep learning methods is achieving results very close to those of human annotators [78], recent work found that extrapolating word sentiment continuous variables based on word embeddings still requires significant work [79].”
Inducing Domain-Specific Sentiment Lexicons from Unlabeled Corpora (2016, background): “Recent work has shown that web-scale sentiment lexicons can be automatically induced for large socially-diffuse domains, such as the internet-at-large (Velikovich et al., 2010) or all of Twitter (Tang et al., 2014).”
Shirtless and Dangerous: Quantifying Linguistic Signals of Gender Bias in an Online Fiction Writing Community (2016, background): “Other researches have crowd-sourced new lexical categories when they do not yet exist (Wagner et al. 2015; Fast, Chen, and Bernstein 2016), an approach we adopt to build categories that capture common gender stereotypes.”
Gender and Representation Bias in GPT-3 Generated Stories (2021, methodology): “Words linked to people via linguistic dependencies can be used to analyze descriptions of people in text (Fast et al., 2016b; Hoyle et al., 2019; Lucy et al., 2020; Bamman et al., 2013; Sap et al., 2017).”