Twitter has become one of the thriving social media platforms since its launch in 2006, while complaints about Twitter bots are constantly on the rise. Since then, many efforts have been devoted to the task of Twitter bot detection and the creation of relevant datasets. However, existing benchmarks generally suffer from low levels of user diversity, limited user information and data scarcity. Therefore, these datasets are not sufficient to train and stably benchmark bot detection measures. To alleviate these problems, we present TwiBot-20, a massive Twitter bot detection benchmark, which contains 229,573 users, 33,488,192 tweets, 8,723,736 user property items and 455,958 follow relationships. TwiBot-20 covers diversified bots and genuine users to better represent the real-world Twittersphere. TwiBot-20 also includes three modals of user information to support both binary classification of single users and community-aware approaches. To the best of our knowledge, TwiBot-20 is the largest Twitter bot detection benchmark to date. We reproduce competitive bot detection methods and conduct a thorough evaluation on TwiBot-20 and two other public datasets. Experiment results demonstrate that existing bot detection measures fail to match their previously claimed performance on TwiBot-20, which suggests that Twitter bot detection remains a challenging task and requires further research efforts.

TwiBot-20: A Comprehensive Twitter Bot Detection Benchmark

Identifying machine learning techniques for classification of target advertising

Twitter has become a vital social media platform while an ample amount of malicious Twitter bots exist and induce undesirable social effects. Successful Twitter bot detection proposals are generally supervised, which rely heavily on large-scale datasets. However, existing benchmarks generally suffer from low levels of user diversity, limited user information and data scarcity. Therefore, these datasets are not sufficient to train and stably benchmark bot detection measures. To alleviate these problems, we present TwiBot-20, a massive Twitter bot detection benchmark, which contains 229,573 users, 33,488,192 tweets, 8,723,736 user property items and 455,958 follow relationships. TwiBot-20 covers diversified bots and genuine users to better represent the real-world Twittersphere. TwiBot-20 also includes three modals of user information to support both binary classification of single users and community-aware approaches. To the best of our knowledge, TwiBot-20 is the largest Twitter bot detection benchmark to date. We reproduce competitive bot detection methods and conduct a thorough evaluation on TwiBot-20 and two other public datasets. Experiment results demonstrate that existing bot detection measures fail to match their previously claimed performance on TwiBot-20, which suggests that Twitter bot detection remains a challenging task and requires further research efforts.

https://dl.acm.org/doi/pdf/10.1145/3459637.3482019

This chapter assesses the varying degrees to which young women in Arab countries are empowered or disempowered, included or excluded. Setting aside the stereotypes of subordination, it presents a nuanced picture
of the changing circumstances of young women today, the struggle against injustice and the triumphs over injustice. The chapter shows how a new generation of young feminists
is challenging the considerable barriers creatively and with determination.

The new dynamics in the inclusion and empowerment of young women. In:The Arab Human Development Report 2016: Youth and the Prospects for Human Development in a Changing Reality

Twitter has become a major social media platform since its launching in 2006, while complaints about bot accounts have increased recently. Although extensive research efforts have been made, the state-of-the-art bot detection methods fall short of generalizability and adaptability. Specifically, previous bot detectors leverage only a small fraction of user information and are often trained on datasets that only cover few types of bots. As a result, they fail to generalize to real-world scenarios on the Twittersphere where different types of bots co-exist. Additionally, bots in Twitter are constantly evolving to evade detection. Previous efforts, although effective once in their context, fail to adapt to new generations of Twitter bots. To address the two challenges of Twitter bot detection, we propose SATAR, a self-supervised representation learning framework of Twitter users, and apply it to the task of bot detection. In particular, SATAR generalizes by jointly leveraging the semantics, property and neighborhood information of a specific user. Meanwhile, SATAR adapts by pre-training on a massive number of self-supervised users and fine-tuning on detailed bot detection scenarios. Extensive experiments demonstrate that SATAR outperforms competitive baselines on different bot detection datasets of varying information completeness and collection time. SATAR is also proved to generalize in real-world scenarios and adapt to evolving generations of social media bots.

SATAR: A Self-supervised Approach to Twitter Account Representation Learning and its Application in Bot Detection

Spam Bots have become a threat to online social networks with their malicious behavior, posting misinformation messages and influencing online platforms to fulfill their motives. As spam bots have become more advanced over time, creating algorithms to identify bots remains an open challenge. Learning low-dimensional embeddings for nodes in graph structured data has proven to be useful in various domains. In this paper, we propose a model based on graph convolutional neural networks (GCNN) for spam bot detection. Our hypothesis is that to better detect spam bots, in addition to defining a features set, the social graph must also be taken into consideration. GCNNs are able to leverage both the features of a node and aggregate the features of a node’s neighborhood. We compare our approach, with two methods that work solely on a features set and on the structure of the graph. To our knowledge, this work is the first attempt of using graph convolutional neural networks in spam bot detection.

Detect Me If You Can: Spam Bot Detection Using Inductive Representation Learning

In order to create an effective article, having great content is essential. However, to achieve this, the writer needs to target a specific audience. A target audience refers to a group of readers that a writer intends to reach with his content. Defining a target audience is substantial because it has a direct effect on adjusting writing style and content of the article. Nowadays, writers rely solely on annotated attributes of articles, such as location and language to understand his/her audience.The aim of this work is to identify the audience attributes of articles, especially not-annotated attributes. Among others, this work focuses on the detection of three key audience attributes of related articles: age, gender, and personality.We compare between multiple machine learning classifiers to detect these attributes. Finally, we demonstrate a prototypical application that enables writers to run existing algorithms such as trend detection and showing related articles that are specific to a defined target audience based on the newly detected attributes.

Identifying Audience Attributes: Predicting Age, Gender and Personality for Enhanced Article Writing

User-generated content on social media platforms is a rich source of latent information about individual variables. Crawling and analyzing this content provides a new approach for enterprises to personalize services and put forward product recommendations. In the past few years, brands made a gradual appearance on social media platforms for advertisement, customers support and public relation purposes and by now it became a necessity throughout all branches. This online identity can be represented as a brand personality that reflects how a brand is perceived by its customers. We exploited recent research in text analysis and personality detection to build an automatic brand personality prediction model on top of the (Five-Factor Model) and (Linguistic Inquiry and Word Count) features extracted from publicly available benchmarks. The proposed model reported significant accuracy in predicting specific personality traits form brands. For evaluating our prediction results on actual brands, we crawled the Facebook API for 100k posts from the most valuable brands' pages in the USA and we visualize exemplars of comparison results and present suggestions for future directions.

Personality Exploration System for Online Social Networks: Facebook Brands As a Use Case

The essence of images users post and share at their social media platform is motivated and elucidated by their individual psychological constructs which are designated as personality traits. In this research, we investigate how social media profile pictures differ based on the personality of the users posting them at their social networking sites. In our experiment, we use profile images from Twitter platform whose personality we predicted based on 1.7 million data points. We conducted our analysis on users faces by extracting unique 50 facial features in order to examine the relationship between personality and profile picture. Our results reveal notable variations in profile picture selection between different personality traits. For example, high to openness males have a high surprise emotion in their profile pictures and a low smile value, while high in openness females have a high happiness emotion in their profile pictures and a high smile value. Finally, various machine learning approaches were investigated to test the effectiveness of these facial features in predicting users' psychological traits. Our results show that training personality models on a granularity based on gender gains higher accuracy. To our knowledge, this work is the first attempt of using ensemble learning methods for personality prediction task from users profile picture.

Facial-Based Personality Prediction Models for Estimating Individuals Private Traits

We demonstrate that easy accessible digital records of behavior such as Facebook Likes can be obtained and utilized to automatically distinguish a wide range of highly delicate personal traits such as the Big Five personality traits. The analysis presented based on a dataset of over 738,000 users conferred their Facebook Likes (95 million unique Like objects), social network activities, posts, egocentric network, demographic characteristics, and results of various self-reported psychometric tests. The proposed model uses a new and unique mapping technique between each Facebook Like object to their corresponding Facebook page category/sub-category object extracted from the API calls as Likes metadata, which is then evaluated as features for a set of machine learning algorithms to predict individual psychodemographic profiles from users Likes. Traditionally, entities where able to access an individual’s personality through having them fill out psychological questionnaires. In this paper, we present a method which indicates that a person’s Big Five personality score can be easily predicted by leveraging the information about the pages a person liked on Facebook.

Raad Bin Tareaf

Papers

Detect Me If You Can: Spam Bot Detection Using Inductive Representation Learning

Identifying Audience Attributes: Predicting Age, Gender and Personality for Enhanced Article Writing

Personality Exploration System for Online Social Networks: Facebook Brands As a Use Case

Facial-Based Personality Prediction Models for Estimating Individuals Private Traits

Towards Automatic Personality Prediction Using Facebook Likes Metadata