Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2201.08387

Understanding and Detecting Hateful Content using Contrastive Learning

Felipe González-Pizarro·Savvas Zannettou
View original source ↗Read source document ↗Export metadata ↓

Abstract

The spread of hate speech and hateful imagery on the Web is a significant problem that needs to be mitigated to improve our Web experience. This work contributes to research efforts to detect and understand hateful content on the Web by undertaking a multimodal analysis of Antisemitism and Islamophobia on 4chan's /pol/ using OpenAI's CLIP. This large pre-trained model uses the Contrastive Learning paradigm. We devise a methodology to identify a set of Antisemitic and Islamophobic hateful textual phrases using Google's Perspective API and manual annotations. Then, we use OpenAI's CLIP to identify images that are highly similar to our Antisemitic/Islamophobic textual phrases. By running our methodology on a dataset that includes 66M posts and 5.8M images shared on 4chan's /pol/ for 18 months, we detect 173K posts containing 21K Antisemitic/Islamophobic images and 246K posts that include 420 hateful phrases. Among other things, we find that we can use OpenAI's CLIP model to detect hateful content with an accuracy score of 0.81 (F1 score = 0.54). By comparing CLIP with two baselines proposed by the literature, we find that CLIP outperforms them, in terms of accuracy, precision, and F1 score, in detecting Antisemitic/Islamophobic images. Also, we find that Antisemitic/Islamophobic imagery is shared in a similar number of posts on 4chan's /pol/ compared to Antisemitic/Islamophobic textual phrases, highlighting the need to design more tools for detecting hateful imagery. Finally, we make available (upon request) a dataset of 246K posts containing 420 Antisemitic/Islamophobic phrases and 21K likely Antisemitic/Islamophobic images (automatically detected by CLIP) that can assist researchers in further understanding Antisemitism and Islamophobia.

Explore related subjects

cs.SIcs.CY

Keep this discovery

BibTeXRIS

Felipe González-Pizarro, Savvas Zannettou. 2022-01-21. Understanding and Detecting Hateful Content using Contrastive Learning. https://arxiv.org/abs/2201.08387

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Link prediction in complex networks via fusing node centrality and local similarity indices

Local similarity indices are widely used in link prediction on complex networks owing to their low computational cost; however, in sparse networks they assign a zero score to every node pair lacking common neighbors, which severely limits their predictive power. A natural remedy is to fuse node centrality indices with local similarity indices: the former provide global importance for the node pair, while the latter capture fine-grained local topology, and the two can be combined into complementary scores within a unified framework. This paper uses PageRank and DomiRank as two representative centrality measures and constructs a centrality--local-similarity fusion framework. The PageRank-based fusion proposed by Charikhi is first generalized to seven classical local similarity indices, and the universality of its improvement is systematically verified on nine real-world network datasets. Furthermore, the DomiRank centrality is introduced to build the DR-MD series of fused indices under a unified weighting coefficient, which overcomes the drawback that the PageRank-based fusion requires index-by-index weight tuning. Results of five-fold cross-validation together with Wilcoxon signed-rank tests show that, under the unified experimental protocol, all DR-MD indices consistently outperform the corresponding local baselines and their PR-MD counterparts on all nine datasets ($p=0.002$), and that the improvements remain robust against perturbations of $\sigma$ and the weighting coefficients within the near-critical parameter plateau; in particular, DR-RA achieves an average AUC of 0.7084, surpassing global methods such as Katz and RWR as well as several advanced similarity indices. The framework is inherently extensible, and its fusion paradigm can be straightforwardly generalized to couple other node centrality indices with local similarity indices.

cs.SI↗
arXiv2026

Chance, Persistent Advantage, and the Generative-AI Era in Open-Source Package Careers

Studies of careers in science, film, music, and books report a common pattern. When a person's most successful work arrives is close to a random draw over the works they produce. How large their successes tend to be, in contrast, follows a stable, person-specific factor. We test whether this pattern holds for open-source software careers and whether it changed when generative AI coding tools arrived. From the complete public record of GitHub push events (2015-2025), we reconstruct 102.2M career works by 6.15M contributors, and for the 908k contributors whose repositories publish packages, we measure each work's impact by how many downstream packages come to depend on it. First, we find that the timing of a career's biggest hit is close to a lottery over their works, as in science and the arts, with a small, replicable lean toward early career that grows as careers get longer. Second, some coders reliably produce higher-impact work than others, but this lasting personal factor accounts for only part of why impact persists (about a fifth in our primary specification); the rest behaves like momentum, success feeding on itself for a period of time. Third, within the same contributors, this structure did not change after ChatGPT's release. The stable factor's weight grew by about as much as it grew for an earlier cohort that simply aged, and subtracting the effect of aging from the effect of generative AI puts the shift at +0.03 (95% CI [-0.22, +0.23]), indistinguishable from zero. The success pattern documented in science and the arts therefore describes open-source careers too, and it shows no detectable break across the arrival of generative AI. These results have implications for how track records on open platforms should be read and on what to expect from generative AI for the careers built on them.

cs.SI↗
arXiv2026

How neighbourhood ideology shapes misinformation belief in densely tied social networks

With the rapid spread of news on social media, understanding the propagation of misinformation is becoming increasingly important. One factor that affects individuals' vulnerability to false information is their ideological predisposition. Despite the large number of agent-based models that focus on social influence as a driver of the spread of false claims, they often fail to explicitly integrate personal ideological biases into belief formation. In this work, we explore how misinformation spreads through the interaction between individuals' ideological biases and social influence. Our model accounts for both the strength of individuals' ideological biases and the extent to which a false claim aligns with their ideology. Social influence modifies the effects of ideological intensity and false claim alignment through network interactions. Notably, the influence of neighbours' ideological intensity on belief is strongly affected by how well those neighbours are connected to one another. These results highlight the importance of considering both network structure and personal ideological biases when modelling misinformation propagation.

cs.SI↗