Online Hate Speech against Women: Automatic Identification of Misogyny and Sexism on Twitter

Frenda, Simona; Ghanem, Bilal; Montes-y-Gómez, Manuel; Rosso, Paolo

doi:10.3233/jifs-179023

Cited by 99 publications

(55 citation statements)

References 17 publications

(29 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Concluding, evaluating all the indexes together, we notice that our dictionarybased algorithm performs very well. In particular, it outperforms the supervised feature-based models of Frenda et al (2019) and of Foucalt et al 2016which report an AC of 76% and a F1 score equal to 72%, respectively. At the same time our method performs more similarly (despite anyway outperforming such methods of more than 8 percentage points) to the computationally intensive deep learning methods implemented in Stowe et al (2019) and Nizzoli et al (2019), which obtain values of the F1 score equal to 83% and 90%, respectively.…”

Section: Table 3 Here (Now At the End Of The File)mentioning

confidence: 81%

“…This estimated classifier is then adopted for predicting the category of a new tweet. This approach is used for example by Frenda et al (2019), who adopt a common method in machine learning as Support Vector Machine (SVM) to automatically detect sexist and misogyny on Twitter. In particular, this study uses, as training dataset, the freely available corpora known as "Automatic Misogyny Identification -IBEREVAL 2018" 5 and considers, as input variables, some lexical and stylistic features of the post.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Dictionary-based Classification of Tweets About Environment

Cameletti¹,

Fabris²,

Schlosser³

et al. 2020

Preprint

View full text Add to dashboard Cite

In the era of social media, the huge availability of digital data (e.g. posts sent through social networks or unstructured data scraped from websites) allows to develop new types of research in a wide range of fields. These types of data are characterized by some advantages such as reduced collection costs, short retrieval times and production of almost real-time outputs. Nevertheless, their collection and analysis can be challenging. For example, particular approaches are required for the selection of posts related to specific topics; moreover, retrieving the information we are interested in inside Twitter posts can be a difficult task.The main aim of this paper is to propose an unsupervised dictionary-based method to filter tweets related to a specific topic, i.e. environment. We start from the tweets sent by a selection of Official Social Accounts clearly linked with the subject of interest. Then, a list of keywords is identified in order to set a topic-oriented dictionary. We test the performance of our method by applying the dictionary to more than 54 million geolocated tweets posted in Great Britain between January and May 2019.

show abstract

Section: Table 3 Here (Now At the End Of The File)mentioning

confidence: 81%

Section: Related Workmentioning

confidence: 99%

Dictionary-based Classification of Tweets About Environment

Cameletti¹,

Fabris²,

Schlosser³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The battle against equality for men and women in cyberspace has intensified online hate speech (Frenda et al, 2019) The amount of hateful material reported in social media against women is even worse than that of their friends. This material offers a model of sexual animosity and prejudices from which AI can benefit.…”

Section: The Consequences Of Gender-biased Artificial Intelligencementioning

confidence: 99%

Using AI Changes the Paradigm of Women's Participation in Politics

Efthymiou

Psomiadi²,

Diareme

et al. 2020

HAPScPBS

View full text Add to dashboard Cite

The effect of AI on how people are viewed and handled in society is important and profound. However, a vicious cycle is maintained with AI's algorithms design and implementation. Among others, predictive models, machine learning and AI algorithms train and test themselves using datasets, as a result, they "learn" mainly based on the data input in a model. Nowadays and in this context, it seems that there is a growing scientific dialogue concerning bias in training AI (Falco, 2019; Lu, 2019; Straw, 2020) as well as whether datasets, on which decisions are made, only represent fractions of reality (Günther et al., 2017). The technology often captures and reproduces regulated and restrictive beliefs regarding gender and race, which are then repetitively strengthened: Gender relations be materialized by inventions and, through their enrolment and incorporation of machinery, masculinity, and femininity gain of turn their importance and character. When robots progress in certain cognitive functions, their comparatively weak abilities will definitely get better. This list incorporates the innovative approach to the dilemma, empathy, negotiation, and belief. Automation and AI will also replace many of today's workers at the same time creating new opportunities for specialized personnel-so that is why women need to get into this emerging sector and ensure that they can secure new jobs when their jobs are squeezed. In addition, AI may provide the ability to alter male and female epistemological assumptions. The narration of "hard" and "soft" intelligence, for instance, is often described as male and female. The rise and development of AI is also seen as pushing economic growth and strengthening political influence. In politics, UK politics still dominates the ambition of economic development by technical advancement. Jude Browne states (Clementine Collett & Sarah Dillon) that a national AI agency equivalent to Human Fertilization and Embryology (HFEA) has yet to be set up by the government of the UK that will fill the divide between national, experts and government, for example. Browne claims that it includes the dominance, primarily guided by the goals of economic wealth, of private interest over the public interest. There is a possibility that economic growth and political influence play an important part in influencing AI laws and policies at the cost of other motivations, which are more morally equal. Consequently, a dual-purpose must be incorporated into an equitable AI policy. Firstly, to ensure there is no rise in social and economic disparity due to the advancement of AI technology. Secondly, to call AI to cut this down. AI must first and foremost enable us to promote our democratic liberties, enhance social harmony, and enhance unity, rather than jeopardise our individual trajectories and networks of solidarity.

show abstract

“…Frenda et al (2019) and ofFoucalt et al (2016) which report an AC of 76% and a F 1 score equal to 72%, respectively. At the same time our method performs more similarly (despite anyway outperforming such methods of more than 8 percentage points) to the computationally intensive deep learning methods implemented inStowe et al (2019) andNizzoli et al (2019), which obtain values of the F 1 score equal to 83% and 90%, respectively.…”

mentioning

confidence: 92%