Social Bias in Elicited Natural Language Inferences
نویسندگان
چکیده
We analyze the Stanford Natural Language Inference (SNLI) corpus in an investigation of bias and stereotyping in NLP data. The human-elicitation protocol employed in the construction of the SNLI makes it prone to amplifying bias and stereotypical associations, which we demonstrate statistically (using pointwise mutual information) and with qualitative examples.
منابع مشابه
A Comparison between Three Methods of Language Sampling: Freeplay, Narrative Speech and Conversation
Objectives: The spontaneous language sample analysis is an important part of the language assessment protocol. Language samples give us useful information about how children use language in the natural situations of daily life. The purpose of this study was to compare Conversation, Freeplay, and narrative speech in aspects of Mean Length of Utterance (MLU), Type-token ratio (TTR), and the numbe...
متن کاملHow Negative Social Bias Affects Memory for Faces: An Electrical Neuroimaging Study
During social interactions, we make inferences about people's personal characteristics based on their appearance. These inferences form a potential prejudice that can positively or negatively bias our interaction with them. Not much is known about the effects of negative bias on face perception and the ability to recognize people faces. This ability was investigated by recording event-related p...
متن کاملItalian Political Communication and Gender Bias: Press Representations of Men/Women Presidents of the Houses of Parliament (1979, 1994, and 2013)
The study considers mass media communication as intertwined with social norms, as assumed by the perspective of social representations. It explores the Italian press communication by focusing on three pairs of men and women politicians with different political orientations and all serving as presidents of the Houses of Parliament in three legislatures. The article concentrates on five newspaper...
متن کاملNatural language processing in mental health applications using non-clinical texts
Natural language processing (NLP) techniques can be used to make inferences about peoples' mental states from what they write on Facebook, Twitter and other social media. These inferences can then be used to create online pathways to direct people to health information and assistance and also to generate personalized interventions. Regrettably, the computational methods used to collect, process...
متن کاملImpaired positive inferential bias in social phobia.
People with social phobia report anticipatory and retrospective judgments about social situations that appear consistent with a negative interpretative bias. However, it is not at all clear that biased interpretative inferences are made "on-line;" that is, at the time that ambiguous information is first encountered. In a previous study, volunteers who were anxious about interviews were found to...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2017