The dramatic increase of the hate speech over the social network sites which leads to an increase in the demands from local and international human rights organizations, research institutions and corporates, legal institutions, religious institutions, independent scholars in social science and universities, social media monitors for controlling and addressing hate speech types by monitoring it and identifying its sources, helping the active institutions in the field of human rights to limit and control this phenomenon and redirect this discourse into interventions from the policy makers which leads to urgent actions and social projects that limits and change the odd social behaviors in societies, mainly Arab societies towards sustainable development, a cohesive society capable of resilience and development. Detecting hate speech on social media is a critical task, the difficulty of automatic detection because social media contain para-linguistic signals (e.g., emoticons and hashtags) as well as a lot of poorly written text in their linguistic content, on the other hand the lack of pre-trained models in Arabic text. Most of the current social media monitoring tools are not based on contextual meaning and require manual efforts from the social media monitors team to do manual filtering especially when monitoring Arabic content.
Our tool, built on different specialized AI technologies like Machine Learning, Natural Language Modeling, and Deep Learning, to enable detection of hate speech based on user defined relevant incidents to automatically find and group tweets that are covering the same event. In this way, all the tweets that are discussing the same topic are grouped into a unit to get a complete coverage of a particular incident or thing that happened. Instead of manually searching for related tweets. Additionally, we compute several properties, such as sentiment score, list of mentioned concepts and topics, number of retweets. Identify relevant risk signals of human rights violations to incite hostility, violence, or discrimination towards a person or a group defined by religion, race, ethnicity, or other factors.
Team:
- Zeina Saadeddin, Technical Specialist
- Diyaa Awad, Technical Specialist
- Amon Alsheikh, Journalist and Digital Media Specialist
- Esraa Awaisa, Statistics specialist
- Zafer Sabbah, Journalist
Supervisors:
- Dr. Abdel-Razzak Natsheh
- Dr. Mohammad Jubran
- Dr. Hamed Abdelhaq