Embeddings-Based Clustering for Target Specific Stances: The Case of a Polarized Turkey

Rashed, Ammar; Kutlu, Mücahid; Darwish, Kareem; Elsayed, Tamer; Bayrak, Cansin

doi:10.48550/arxiv.2005.09649

Cited by 11 publications

(15 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The user embeddings are projected to a two dimensional space using UMAP (McInnes, Healy, and Melville 2020) and clustered using hierarchical density based clustering (HDB-SCAN) (McInnes and Healy 2017). According to Rashed et al (2020), this method results in clusters indicative of users' stance on the specific event.…”

Section: User Embeddingsmentioning

confidence: 99%

“…Our section on categorical analysis further explains the distinctiveness of the COVID-19 clustering. Figure 3 shows the wordclouds, based on prominence scores (Rashed et al 2020), for each of the stances and helps underscore the differences. On Article 370 and CAA/NRC, the INC side focuses on the act being "undemocratic" while the pro-BJP tweets celebrate them as "historic", having citizen support, and highlighting Hindu or refugee-related issues.…”

Section: Stance Detection With Polarized Clustersmentioning

confidence: 99%

See 1 more Smart Citation

Divided We Rule: Influencer Polarization on Twitter During Political Crises in India

Dash¹,

Mishra²,

Shekhawat³

et al. 2021

Preprint

View full text Add to dashboard Cite

Influencers are key to the nature and networks of information propagation on social media. Influencers are particularly important in political discourse through their engagement with issues, and may derive their legitimacy either solely or partly through online operation, or have an offline sphere of expertise such as entertainers, journalists etc. To quantify influencers' political engagement and polarity, we use Google's Universal Sentence Encoder (USE) to encode the tweets of 6k influencers and 26k Indian politicians during political crises in India. We then obtain aggregate vector representations of the influencers based on their tweet embeddings, which alongside retweet graphs help compute the stance and polarity of these influencers with respect to the political issues. We find that while on COVID-19 there is a confluence of influencers on the side of the government, on three other contentious issues around citizenship, Kashmir's statehood, and farmers' protests, it is mainly government-aligned fan accounts that amplify the incumbent's positions. We propose that this method offers insight into the political schisms in present-day India, but also offers a means to study influencers and polarization in other contexts.

show abstract

Section: User Embeddingsmentioning

confidence: 99%

Section: Stance Detection With Polarized Clustersmentioning

confidence: 99%

Divided We Rule: Influencer Polarization on Twitter During Political Crises in India

Dash¹,

Mishra²,

Shekhawat³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Finally, over 43M English tweets are collected 2 from these 26k politicians and 6k influencers accounts between June 2019 -March 2021. The tweets are preprocessed in a manner similar to Rashed et al [51], including case-folding and removal of links, emojis, punctuations and non alphanumeric characters.…”

Section: Data Collectionmentioning

confidence: 99%

Insights Into Incitement: A Computational Perspective on Dangerous Speech on Twitter in India

Dash¹,

Grover²,

Shekhawat³

et al. 2021

Preprint

View full text Add to dashboard Cite

Dangerous speech on social media platforms can be framed as blatantly inflammatory, or be couched in innuendo. It is also centrally tied to who engages it -it can be driven by openly sectarian social media accounts, or through subtle nudges by influential accounts, allowing for complex means of reinforcing vilification of marginalized groups, an increasingly significant problem in the media environment in the Global South. We identify dangerous speech by influential accounts on Twitter in India around three key events, examining both the language and networks of messaging that condones or actively promotes violence against vulnerable groups. We characterize dangerous speech users by assigning Danger Amplification Belief scores and show that dangerous users are more active on Twitter as compared to other users as well as most influential in the network, in terms of a larger following as well as volume of verified accounts. We find that dangerous users have a more polarized viewership, suggesting that their audience is more susceptible to incitement. Using a mix of network centrality measures and qualitative analysis, we find that most dangerous accounts tend to either be in mass media related occupations or allied with low-ranking, right-leaning politicians, and act as "broadcasters" in the network, where they are best positioned to spearhead the rapid dissemination of dangerous speech across the platform.

show abstract

“…With advances in deep neural networks, recent work applied deep learning models to detect polarity by mapping people's systems of belief into a latent space [20,33,37,44,46,49]. For example, Jiang et al [20] developed a weakly supervised model, Retweet-BERT, to predict the polarity of users on Twitter based on network structures and content features.…”

Section: Related Workmentioning

confidence: 99%

Unsupervised Belief Representation Learning with Information-Theoretic Variational Graph Auto-Encoders

Li,

Shao,

Sun

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper develops a novel unsupervised algorithm for belief representation learning in polarized networks that (i) uncovers the latent dimensions of the underlying belief space and (ii) jointly embeds users and content items (that they interact with) into that space in a manner that facilitates a number of downstream tasks, such as stance detection, stance prediction, and ideology mapping.Inspired by total correlation in information theory, we propose a novel Information-Theoretic Variational Graph Auto-Encoder (In-foVGAE) that learns to project both users and content items (e.g., posts that represent user views) into an appropriate disentangled latent space. In order to better disentangle orthogonal latent variables in that space, we develop total correlation regularization, PI control module, and adopt rectified Gaussian Distribution for the latent space. The latent representation of users and content can then be used to quantify their ideological leaning and detect/predict their stances on issues. We evaluate the performance of the proposed InfoVGAE on three real-world datasets, of which two are collected from Twitter and one from U.S. Congress voting records. The evaluation results show that our model outperforms state-of-the-art unsupervised models and produce comparable result with supervised models. We also discuss stance prediction and user ranking within ideological groups. The code and datasets are available at https://github.com/happydeadline/InfoVGAE.

show abstract

Embeddings-Based Clustering for Target Specific Stances: The Case of a Polarized Turkey

Cited by 11 publications

References 27 publications

Divided We Rule: Influencer Polarization on Twitter During Political Crises in India

Divided We Rule: Influencer Polarization on Twitter During Political Crises in India

Insights Into Incitement: A Computational Perspective on Dangerous Speech on Twitter in India

Unsupervised Belief Representation Learning with Information-Theoretic Variational Graph Auto-Encoders

Contact Info

Product

Resources

About