Online Library

With the spread of social networks and their unfortunate use for hate speech, automatic detection of the latter has become a pressing problem. In this paper, we reproduce seven state-of-the-art hate speech detection models from prior work, and show that they perform well only when tested on the same type of data they were trained on. Based on these results, we argue that for successful hate speech detection, model architecture is less important than the type of data and labeling criteria. We further show that all proposed detection techniques are brittle against adversaries who can (automatically) insert typos, change word boundaries or add innocuous words to the original hate speech. A combination of these methods is also effective against Google Perspective – a cuttingedge solution from industry. Our experiments demonstrate that adversarial training does not completely mitigate the attacks, and using character-level features makes the models systematically more
attack-resistant than using word-level features.

To prevent and counter the spread of illegal hate speech online, in May 2016, the Commission agreed with Facebook, Microsoft, Twitter and YouTube a “Code of conduct on countering illegal hate speech online”. The implementation of the Code of Conduct is evaluated through a regular monitoring exercise set up in collaboration with a network of organisations located in the different EU countries. Using a commonly agreed methodology, these organisations test how the IT companies are implementing the commitments in the Code.

This is an analysis of what the Alt-Right stands for and who its followers and leaders are. Including exclusive interviews with members of the movement and evidence linking extremists with terror attacks and hate crimes, it is clear that despite its high-profile support, the movement’s lack of a coherent base and its contradictory tendencies is already leading to its downfall.

In recent years, the increasing propagation of hate speech on social media and the urgent need for effective counter-measures have drawn significant investment from governments, companies, and empirical research. Despite a large number of emerging scientific studies to address the problem, a major limitation of existing work is the lack of comparative evaluations, which makes it difficult to assess the contribution of individual works. This paper introduces a new method based on a deep neural network combining convolutional and gated recurrent networks. We conduct an extensive evaluation of the method against several baselines and state of the art on the largest collection of publicly available Twitter datasets to date, and show that compared to previously reported results on these datasets, our proposed method is able to capture both word sequence and order information in short texts, and it sets new benchmark by outperforming on 6 out of 7 datasets by between 1 and 13% in F1. We also extend the existing dataset collection on this task by creating a new dataset covering different topics.

The paper investigates the potential effects user features have on hate speech classification. A quantitative analysis of Twitter data was conducted to better understand user characteristics, but no correlations were found between hateful text and the characteristics of the users who had posted it. However, experiments with a hate speech classifier based on datasets from three different languages showed that combining certain user features with textual features gave slight improvements of classification performance. While the incorporation of user features resulted in varying impact on performance for the different datasets used, user network-related features provided the most consistent improvements.

The paper describes a recently-created Twitter corpus of about 6,000 tweets, annotated for hate speech against immigrants, and developed to be a reference dataset for an automatic system of hate speech monitoring. The annotation scheme was therefore specifically designed to account for the multiplicity of factors that can contribute to the definition of a hate speech notion, and to offer a broader tagset capable of better representing all those factors, which may increase, or rather mitigate, the impact of the message. This resulted in a scheme that includes, besides hate speech, the following categories: aggressiveness, offensiveness, irony, stereotype, and (on an experimental basis) intensity. The paper hereby presented namely focuses on how this annotation scheme was designed and applied to the corpus. In particular, also comparing the annotation produced by CrowdFlower contributors and by expert annotators, we make some remarks about the value of the novel resource as gold standard, which stems from a preliminary qualitative analysis of the annotated data and on future corpus development.