Machine learning models have had discernible achievements in a myriad of applications. However, most these are black-boxes, and it is obscure how the decisions made by them. This makes unreliable untrustworthy. To provide insights into decision making processes models, variety traditional interpretable been proposed. Moreover, to generate more humanfriendly explanations, recent work on interpretability tries answer questions related causality such as "Why does this model decisions?" or "Was...

Data annotation is the labeling or tagging of raw data with relevant information, essential for improving efficacy machine learning models. The process, however, labor-intensive and expensive. emergence advanced Large Language Models (LLMs), exemplified by GPT-4, presents an unprecedented opportunity to revolutionize automate intricate process annotation. While existing surveys have extensively covered LLM architecture, training, general applications, this paper uniquely focuses on their...

Abstract The creation, dissemination, and consumption of disinformation fabricated content on social media is a growing concern, especially with the ease access to such sources, lack awareness existence false information. In this article, we present an overview techniques explored date for combating various forms. We introduce different forms disinformation, discuss factors related spread elaborate inherent challenges in detecting show some approaches mitigating via education, research,...

With social media being a major force in information consumption, accelerated propagation of fake news has presented new challenges for platforms to distinguish between legitimate and news. Effective detection is non-trivial task due the diverse nature domains expensive annotation costs. In this work, we address limitations existing automated models by incorporating auxiliary (e.g., user comments user-news interactions) into novel reinforcement learning-based model called REinforced Adaptive...

The rise of fake news in the past decade has brought with it a host consequences, from swaying opinions on elections to generating uncertainty during pandemic. A majority methods developed combat disinformation either focus content or malicious actors who generate it. However, virality is largely dependent upon users propagate deeper understanding these can contribute development framework for identifying are likely spread news. In this work, we study characteristics and motivational factors...

Echo chambers on social media are a significant problem that can elicit number of negative consequences, most recently affecting the response to COVID-19. promote conspiracy theories about virus and found be linked vaccine hesitancy, less compliance with mask mandates, practice distancing. Moreover, echo is connected other pertinent issues like political polarization spread misinformation. An chamber defined as network users in which only interact opinions support their pre-existing beliefs...

Echo chambers may exclude social media users from being exposed to other opinions, therefore, can cause rampant negative effects. Among abundant evidence are the 2016 and 2020 US presidential elections conspiracy theories polarization, as well COVID-19 disinfodemic. To help better detect echo mitigate its effects, this paper explores mechanisms attributes of in media. In particular, we first illustrate four primary related three main factors: human psychology, networks, automatic systems. We...

It is crucial to detect and manage stress as early possible before it becomes a severe mental physical health problem. Some authors even introduce “silent killer” emphasize the significance of management. Traumatic global events such COVID-19 have amplified throughout online communities quite common see that social media users often vent about their problems or situations online. The ability person's from posts on platforms like Reddit Twitter in timely manner can help management...

We propose a generative Causal Adversarial Network (CAN) for learning and sampling from conditional interventional distributions. In contrast to the existing CausalGAN which requires causal graph be given, our proposed framework learns relations data generates samples accordingly. The CAN comprises two-fold process namely Label Generation (LGN) Conditional Image (CIGN). LGN is GAN-based architecture model over labels. sampled labels are then fed CIGN, GAN architecture, relationships amongst...

An Echo Chamber on social media refers to the environment where like-minded people hear echo of each others' voices, opinions, or beliefs, which reinforce their own. Chambers can turn platforms into collaborative venues that polarize and radicalize users rather than broadening exposure diverse information. Having a quantified metric for measuring effect aid moderators policymakers in tracking mitigating online polarization radicalization. Existing methods detection are either...

Machine learning models have had discernible achievements in a myriad of applications. However, most these are black-boxes, and it is obscure how the decisions made by them. This makes unreliable untrustworthy. To provide insights into decision making processes models, variety traditional interpretable been proposed. Moreover, to generate more human-friendly explanations, recent work on interpretability tries answer questions related causality such as "Why does this model decisions?" or "Was...

Neural network-based embeddings have been the mainstream approach for creating a vector representation of text to capture lexical and semantic similarities dissimilarities. In general, existing encoding methods dismiss punctuation as insignificant information; consequently, they are routinely treated predefined token/word or eliminated in pre-processing phase. However, could play significant role semantics sentences, "Let's eat\hl{,} grandma" eat grandma". We hypothesize that...

The creation, dissemination, and consumption of disinformation fabricated content on social media is a growing concern, especially with the ease access to such sources, lack awareness existence false information. In this paper, we present an overview techniques explored date for combating various forms. We introduce different forms disinformation, discuss factors related spread elaborate inherent challenges in detecting show some approaches mitigating via education, research, collaboration....

Online Social Networks (OSNs) facilitate access to a variety of data allowing researchers analyze users' behavior and develop user behavioral analysis models. These models rely heavily on the observed which is usually biased due participation inequality. This inequality consists three groups online users: lurkers - users that solely consume content, engagers contribute little content creation, contributors are responsible for creating majority content. Failing consider contribution all while...

Social media platforms provide a rich environment for analyzing user behavior. Recently, deep learning-based methods have been mainstream approach social analysis models involving complex patterns. However, these are susceptible to biases in the training data, such as participation inequality. Basically, mere 1% of users generate majority content on networking sites, while remaining users, though engaged varying degrees, tend be less active creation and largely silent. These silent consume...