Homophobic speech is a form of hate speech. Social media enables hate speech to spread rapidly and widely through the internet, and unlike offline hate speech, can persist indefinitely, thereby prolonging its impact. ...
详细信息
ISBN:
(纸本)9783031785405;9783031785412
Homophobic speech is a form of hate speech. Social media enables hate speech to spread rapidly and widely through the internet, and unlike offline hate speech, can persist indefinitely, thereby prolonging its impact. Due to the adverse impact of hate speech, policymakers have called for greater action from online platforms to moderate and remove hate speech, including homophobic content. While homophobic hate speech is prevalent in online soccer discourses, there are few studies on this empirical context in general and specifically on the use of Large Language Models (LLMs) for detecting such speech. This study addresses this gap by proposing a homophobic speech text classification pipeline. We introduce H-DICT, a new general dictionary for identifying potential homophobic content in documents, and leverage this dictionary to curate and manually label an annotated dataset of homophobic and non-homophobic samples from the UEFA European Football Championships (the Euros) discourse on Twitter. We fine-tune and evaluate five large language models (LLMs) based on the BERT architecture BERT, DistilBERT, RoBERTa, BERT Hate, and RoBERTa Offensive - and use Integrated Gradients, an explainable AI technique to explain each model's predictions. RoBERTa Offensive, an LLM fine-tuned specifically for detecting offensive language, presented the best performance when compared to the other LLMs.
Biased information on social media significantly influences public perception by reinforcing stereotypes and deepening societal divisions. Previous research has often isolated specific bias dimensions, such as politic...
详细信息
ISBN:
(纸本)9783031785375;9783031785382
Biased information on social media significantly influences public perception by reinforcing stereotypes and deepening societal divisions. Previous research has often isolated specific bias dimensions, such as political or racial bias, without considering their interrelationships across different domains. The dynamic nature of social media, with its shifting user behaviors and trends, further challenges the efficacy of existing benchmarks. Addressing these gaps, our research introduces a novel dataset derived from five years of YouTube comments, annotated for a wide range of biases including gender, race, politics, and hate speech. This dataset covers diverse areas such as politics, sports, healthcare, education, and entertainment, revealing complex bias interplays. Through detailed statistical analysis, we identify distinct bias expression patterns and intra-domain correlations, setting the stage for developing systems that detect multiple biases concurrently. Our work enhances media bias identification and contributes to the creation of tools for fairer social media consumption.
In today's systems, privacy is often at odds with utility: users that reveal little information about themselves get restricted functionality, and service providers mistrust them. In practice, systems tip to eithe...
详细信息
ISBN:
(纸本)9783031786785;9783031786792
In today's systems, privacy is often at odds with utility: users that reveal little information about themselves get restricted functionality, and service providers mistrust them. In practice, systems tip to either full anonymity (e.g. Monero), or full utility (e.g. Bitcoin). Well-known cryptographic primitives for bridging this gap exist: anonymous credentials (AC) let users disclose a subset of their credentials' attributes, revealing to service providers "just what they need";group signatures (GS) allow users to authenticate anonymously, to be de-anonymized "just when deemed necessary". However, these primitives are hard to deploy. Current AC and GS variants reach specific points in the privacy-utility tradeoff, which we point as counter-productive engineering-wise, as it requires full and error-prone re-engineering to adjust the tradeoff. Also, so far, GS and AC have been studied separately by theoretical research. We take the first steps toward unifying and generalizing both domains, with the goal of bringing their benefits to practice, in a flexible way. We give a common model capturing their core properties, and use functional placeholders to subsume intermediate instantiations of the privacy-utility tradeoff under the same model. To prove its flexibility, we show how concrete variants of GS, AC (and others, like ring signatures) can be seen as special cases of our scheme - to which we refer as universal anonymous signatures (UAS). In practice, this means that instantiations following our construction can be configured to behave as variant X of a GS scheme, or as variant Y of an AC scheme, by tweaking a few functions.
Provoked by numerous art-less walks during the pandemic, our Open Air Gallery AR project, is a smartphone-based locative media platform to display digital galleries and music albums with gamified engagement. This shor...
详细信息
ISBN:
(纸本)9783031784521;9783031784538
Provoked by numerous art-less walks during the pandemic, our Open Air Gallery AR project, is a smartphone-based locative media platform to display digital galleries and music albums with gamified engagement. This short work-in-progress paper presents one such "gallery game," a new type of interactive multimedia for virtual exhibition in nature or a park. Geri Hahn Art: Living Out Loud AR (2024) features the art of an artist-synesthete who uses synesthesia as her creative inspiration. The app includes music by Svetlana Rudenko (Piano, Logic Pro) and Mads Haahr (Synths), composed to transfer the art expression into music language using a unique multisensory design approach where art is translated into music, and location is chosen co-create the experience. Recorded talks with the artist give insights into her life in art, her inspiration and craft of expression. The gallery game genre allows access for a wide audience without the limitations of a physical exhibition venue (e.g., availability, costs for rent). Our preliminary results show a high level of interest in the genre as well as promising social benefits of cooperative play.
On 15th September 2022, The Merge marked the Ethereum network's transition from computation-hardness-based consensus (proof-of-work) to a committee-based consensus mechanism (proof-of-stake). As a result, all the ...
详细信息
ISBN:
(纸本)9783031692307;9783031692314
On 15th September 2022, The Merge marked the Ethereum network's transition from computation-hardness-based consensus (proof-of-work) to a committee-based consensus mechanism (proof-of-stake). As a result, all the specialized hardware and GPUs that were being used by miners ceased to be profitable in the main Ethereum network. Miners were then left with the decision of how to re-purpose their hardware. One such choice was to try and make a profit mining another existing PoW system. In this study, we explore this choice by analyzing the hashrate increase in the top PoW networks following the merge. Our findings reveal that the peak increase in hashrate to other PoW networks following The Merge represents an adoption of at least 41% of the hashrate that was present in Ethereum, with 30% thereof remaining over 5 months later. Though we measure a drastic decrease in profitability by almost an order of magnitude, the continued presence of miners halts claims that power consumption was instantly addressed by Ethereum's switch to PoS.
The prevalence of offensive content on the internet, encompassing hate speech and cyberbullying, is a pervasive issue worldwide. Consequently, it has garnered significant attention from the machine learning (ML) and n...
详细信息
ISBN:
(纸本)9783031785405;9783031785412
The prevalence of offensive content on the internet, encompassing hate speech and cyberbullying, is a pervasive issue worldwide. Consequently, it has garnered significant attention from the machine learning (ML) and natural language processing (NLP) communities. As a result, numerous systems have been developed to automatically identify potentially harmful content and to mitigate its impact. These systems can follow two approaches;(i) Use publicly available models and application endpoints, including prompting large language models (LLMs) (ii) Annotate datasets and train ML models on them. However, both approaches lack an understanding of how generalizable they are. Furthermore, the applicability of these systems is often questioned in off-domain and practical environments. This paper empirically evaluates the generalizability of offensive language detection models and datasets across a novel generalized benchmark: GenOffense. We answer three research questions on generalizability. Our findings will be useful in creating robust real-world offensive language detection systems.
The public page is a popular online social community platform. These pages form a network by liking each other. Location classification of public pages has been studied at the country and state levels. In this paper, ...
详细信息
ISBN:
(纸本)9783031785535;9783031785542
The public page is a popular online social community platform. These pages form a network by liking each other. Location classification of public pages has been studied at the country and state levels. In this paper, we explored the task of public page classification by cities within California. We introduced a virtual geographic structure for city clusters resembling counties in California. We developed a clustering algorithm that leverages the confusion matrix from flat city classification to construct the virtual geographic city structure. Then, adopting a two-stage hierarchical classification strategy-first classifying pages by city cluster and then within clusters by city-we enhanced the accuracy from 0.6928 of flat city classification to 0.8014.
Complex DeFi services are usually constructed by composing a variety of simpler smart contracts. The permissionless nature of the blockchains where these smart contracts are executed makes DeFi services exposed to sec...
详细信息
ISBN:
(纸本)9783031786785;9783031786792
Complex DeFi services are usually constructed by composing a variety of simpler smart contracts. The permissionless nature of the blockchains where these smart contracts are executed makes DeFi services exposed to security risks, since adversaries can target any of the underlying contracts to economically damage the compound service. We introduce a new notion of secure composability of smart contracts, which ensures that adversaries cannot economically harm the compound contract by interfering with its dependencies.
According to the Pew Research Center, a majority of X (formerly Twitter) users in the U.S. (55%) regularly consume news through the platform, exceeding the ratio for all other major social media platforms. Still, the ...
详细信息
ISBN:
(纸本)9783031785474;9783031785481
According to the Pew Research Center, a majority of X (formerly Twitter) users in the U.S. (55%) regularly consume news through the platform, exceeding the ratio for all other major social media platforms. Still, the current literature falls short in providing insights into the relative interaction patterns seen for different classes of news on this platform. To address this gap, this study provides a large-scale analysis of user interactions with different news classes, emphasizing both the bias and reliability of the publishers. To this end, we have compiled a robust dataset comprising more than 75 million tweets posted over 56 months by 2,041 labeled U.S. news publishers. Using this dataset, we study the engagement patterns across news categories, identifying several statistically significant variances. Understanding these dynamics is crucial for developing informed strategies for news dissemination, audience targeting, and content moderation. Accordingly, this study offers data-driven insights to support such strategy development.
This study enhances stance detection on social media by incorporating deeper psychological attributes, specifically individuals' moral foundations. These theoretically-derived dimensions aim to provide an interpre...
详细信息
ISBN:
(纸本)9783031785405;9783031785412
This study enhances stance detection on social media by incorporating deeper psychological attributes, specifically individuals' moral foundations. These theoretically-derived dimensions aim to provide an interpretable profile of an individual's moral concerns which, in recent work, has been linked to behaviour in a range of domains including society, politics, health, and the environment. In this paper, we investigate how moral foundation dimensions can contribute to detecting an individual's stance on a given target. Specifically, we incorporate moral foundation features extracted from text, along with semantic features, to classify stances at both message- and user-levels using traditional machine learning and Large Language Models (LLMs). Our preliminary results suggest that encoding moral foundations can enhance the performance of stance detection tasks, but with notable heterogeneity across task type, models, and datasets. In addition, we illustrate meaningful associations between specific moral foundations and online stances on target topics. The findings from this study highlight the importance of considering deeper psychological attributes in stance classification tasks, and underscore the role of moral foundations in guiding online social behavior.
暂无评论