Association for the Advancement of Artificial Intelligence: AAAI Publications
Not a member yet
26155 research outputs found
Sort by
FedVCK: Non-IID Robust and Communication-Efficient Federated Learning via Valuable Condensed Knowledge for Medical Image Analysis
No full textFederated learning has become a promising solution for collaboration among medical institutions. However, data owned by each institution would be highly heterogeneous and the distribution is always non-independent and identical distribution (non-IID), resulting in client drift and unsatisfactory performance. Despite existing federated learning methods attempting to solve the non-IID problems, they still show marginal advantages but rely on frequent communication which would incur high costs and privacy concerns. In this paper, we propose a novel federated learning method: Federated learning via Valuable Condensed Knowledge (FedVCK). We enhance the quality of condensed knowledge and select the most necessary knowledge guided by models, to tackle the non-IID problem within limited communication budgets effectively. Specifically, on the client side, we condense the knowledge of each client into a small dataset and further enhance the condensation procedure with latent distribution constraints, facilitating the effective capture of high-quality knowledge. During each round, we specifically target and condense knowledge that has not been assimilated by the current model, thereby preventing unnecessary repetition of homogeneous knowledge and minimizing the frequency of communications required. On the server side, we propose relational supervised contrastive learning to provide more supervision signals to aid the global model updating. Comprehensive experiments across various medical tasks show that FedVCK can outperform state-of-the-art methods, demonstrating that it's non-IID robust and communication-efficient
A Novel Diffusion Model for Pairwise Geoscience Data Generation with Unbalanced Training Dataset
No full textRecently, the advent of generative AI technologies has made transformational impacts on our daily lives, yet its application in scientific applications remains in its early stages. Data scarcity is a major, well-known barrier in data-driven scientific computing, so physics-guided generative AI holds significant promise. In scientific computing, most tasks study the conversion of multiple data modalities to describe physical phenomena, for example, spatial and waveform in seismic imaging, time and frequency in signal processing, and temporal and spectral in climate modeling; as such, multi-modal pairwise data generation is highly required instead of single-modal data generation, which is usually used in natural images (e.g., faces, scenery). Moreover, in real-world applications, the unbalance of available data in terms of modalities commonly exists; for example, the spatial data (i.e., velocity maps) in seismic imaging can be easily simulated, but real-world seismic waveform is largely lacking. While the most recent efforts enable the powerful diffusion model to generate multi-modal data, how to leverage the unbalanced available data is still unclear. In this work, we use seismic imaging in subsurface geophysics as a vehicle to present "UB-Diff", a novel diffusion model for multi-modal paired scientific data generation. One major innovation is a one-in-two-out encoder-decoder network structure, which can ensure pairwise data is obtained from a co-latent representation. Then, the co-latent representation will be used by the diffusion process for pairwise data generation. Experimental results on the OpenFWI dataset show that UB-Diff significantly outperforms existing techniques in terms of Fréchet Inception Distance (FID) score and pairwise evaluation, indicating the generation of reliable and useful multi-modal pairwise data
What's in a Prompt?: A Large-Scale Experiment to Assess the Impact of Prompt Design on the Compliance and Accuracy of LLM-Generated Text Annotations
No full textManually annotating data for computational social science tasks can be costly, time-consuming, and emotionally draining. While recent work suggests that LLMs can perform such annotation tasks in zero-shot settings, little is known about how prompt design impacts LLMs' compliance and accuracy. We conduct a large-scale multi-prompt experiment to test how model selection (GPT-4o, GPT-3.5, PaLM2, and Falcon7b) and prompt design features (definition inclusion, output type, explanation, and prompt length) impact the compliance and accuracy of LLM-generated annotations on four highly relevant and diverse CSS tasks (toxicity, sentiment, rumor stance, and news frames). Our results show that LLM compliance and accuracy are prompt-dependent. For instance, prompting for numerical scores instead of labels reduces all LLMs' compliance and accuracy. Concise prompts can significantly reduce prompting costs but also lead to lower accuracy on tasks like toxicity. Furthermore, minor prompt changes like asking for an explanation can cause large changes in the distribution of LLM-generated labels. By assessing the impact of prompt design on the quality and distribution of LLM-generated annotations, this work serves as both a practical guide and a warning for using LLMs in CSS research
Clustering Internet Memes Through Template Matching and Multi-Dimensional Similarity
No full textMeme clustering is critical for toxicity detection, virality modeling, and typing, but it has received little attention in previous research. Clustering similar Internet memes is challenging due to their multimodality, cultural context, and adaptability. Existing approaches rely on databases, overlook semantics, and struggle to handle diverse dimensions of similarity. This paper introduces a novel method that uses template-based matching with multi-dimensional similarity features, thus eliminating the need for predefined databases and supporting adaptive matching. Memes are clustered using local and global features across similarity categories such as form, visual content, text, and identity. Our combined approach outperforms existing clustering methods, producing more consistent and coherent clusters, while similarity-based feature sets enable adaptability and align with human intuition. We make all supporting code publicly available to support subsequent research
Cross-Partisan Interactions on Twitter
No full textMany social media studies argue that social media creates echo chambers where some users only interact with peers of the same political orientation. However, recent studies suggest that a substantial amount of Cross-Partisan Interactions (CPIs) do exist --- even within echo chambers, but they may be toxic. There is no consensus about how such interactions occur and when they lead to healthy or toxic dialogue. In this paper, we study a comprehensive Twitter dataset that consists of 3 million tweets from 2020 related to the U.S. context to understand the dynamics behind CPIs. We investigate factors that are more associated with such interactions, including how users engage in CPIs, which topics are more contentious, and what are the stances associated with healthy interactions. We find that CPIs are significantly influenced by the nature of the topics being discussed, with politically charged events acting as strong catalysts. The political discourse and pre-established political views sway how users participate in CPIs, but the direction in which users go is nuanced. While Democrats engage in cross-partisan interactions slightly more frequently, these interactions often involve more negative and nonconstructive stances compared to their intra-party interactions. In contrast, Republicans tend to maintain a more consistent tone across interactions. Although users are more likely to engage in CPIs with popular accounts in general, this is less common among Republicans who often engage in CPIs with accounts with a low number of followers for personal matters. Our study has implications beyond Twitter as identifying topics with low toxicity and high CPI can help highlight potential opportunities for reducing polarization while topics with high toxicity and low CPI may action targeted interventions when moderating harm
A Growing Sense of Alienation: Spirals of Silence and Suppression of Structural Circumstances of Suicide in News
No full textSuicide is a leading cause of death in the United States. Global safe reporting guidelines for news reports of suicide intend to mitigate associations of increased suicide incidence and stigma. However, recent research suggests more latent patterns in news beyond the guidelines could still contribute to suicide outcomes such as inhibited help-seeking and isolation. Using the Theory of Spiral of Silence to center isolation, we take a mixed-methods approach to analyze 22,021 articles (2020-2024) and use a zero-shot learning large language model (LLM) classifier to detect suppression of four structural circumstances of suicide: financial/job, legal, school, and access to physical/mental healthcare. We find that circumstance disclosure by news publishers diverges by political leaning, financial (p = 0.016), legal (p < 0.001), and school (p < 0.001); and by regionality, legal (p < 0.001) and health (p < 0.001). We qualify mechanisms of suppression using topic modeling and content sharing networks (CSNs). The spiral of silence lens highlights that left leaning publishers are more likely to disclose systemically or socially collective circumstances. In contrast, right leaning outlets suppress those and instead disclose instances that blame individuals for their experiences. Our work highlights how news reporting can downplay structural factors contributing to suicide. Content Warning: This paper discusses suicide deaths reported in news articles and may be sensitive to readers
Hot Tweets and Cold Posts: Variation in US Congresspeople's Ideological Presentation on Twitter and Facebook Over Time
No full textThis work presents a novel observational study of US congresspeople’s link-based news-sharing behaviors and ideological presentations across Facebook and Twitter. By analyzing the web domains these politicians share, we estimate their political ideologies and measure ideological extremity across the political and social contexts of these platforms. Our findings show that these politicians present as more ideologically extreme on Facebook than they appear on Twitter, particularly among Democrats. However, this difference is relatively small compared to the ideological shift between a politician’s publicly funded official account and their campaign account—a shift that is roughly seven times larger. Finally, we observe that these changes are not uniform over time across parties; expressed polarization within the Democratic Party notably increased from 2013 to 2017 before stabilizing, while the Republican Party became markedly more polarized starting in 2020. While more research is needed to identify the specific affordances that contribute to more expressed polarization on Facebook and potential temporal dynamics between these platforms, this work highlights the limitations of studies that focus on single platforms and opens new avenues for future research into how differences across online social spaces may impact political polarization
Unifying the Extremes: Developing a Unified Model for Detecting and Predicting Extremist Traits and Radicalization
No full textThe proliferation of ideological movements into extremist factions via social media has become a global concern. While radicalization has been studied extensively within the context of specific ideologies, our ability to accurately characterize extremism in more generalizable terms remains underdeveloped. In this paper, we propose a novel method for extracting and analyzing extremist discourse across a range of online ideological community forums. By focusing on verbal behavioral signatures of extremist traits, we develop a framework for quantifying extremism at both user and community levels. Our research identifies 11 distinct factors, which we term "The Extremist Eleven," as a generalized psychosocial model of extremism. Applying our method to various online communities, we demonstrate an ability to characterize ideologically diverse communities across the 11 extremist traits. We demonstrate the power of this method by analyzing user histories from members of the incel community. We find that our framework accurately predicts which users join the incel community up to 10 months before their actual entry with an AUC of > 0.6, steadily increasing to AUC ~ 0.9 three to four months before the event. Further, we find that upon entry into an ideological forum, the users tend to maintain their level of extremist traits within the community, while still remaining distinguishable from the general online discourse. Our findings contribute to the study of extremism by introducing a more holistic, cross-ideological approach that transcends traditional, trait-specific models
Large-Scale Analysis of Online Questions Related to Opioid Use Disorder on Reddit
No full textOpioid use disorder (OUD) is a leading health problem that affects individual well-being as well as general public health. Due to a variety of reasons including stigma faced by people who use opioids, online communities for recovery and support were formed on different social media platforms. In these communities, people share their experiences and solicit information by asking questions to learn about opioid use and recovery. These communities however do not always contain clinically verified information. In this paper, we study natural language questions asked in the context of OUD-related discourse on Reddit. We adopt transformer-based question detection along with hierarchical clustering across 19 subreddits to identify six coarse-grained categories and 69 fine-grained categories of OUD-related questions. Our analysis uncovers ten areas of information-seeking from Reddit users in the context of OUD: drug sales, specific drug-related questions, OUD treatment, drug uses, side effects, withdrawal, lifestyle, drug testing, pain management and others, during the study period of 2018-2021. Our work provides a major step in improving the understanding of OUD-related questions people ask unobtrusively on Reddit. We finally discuss technological interventions and public health harm reduction techniques based on the topics of these questions
The Susceptibility Paradox in Online Social Influence
No full textUnderstanding susceptibility to online influence is crucial for mitigating the spread of misinformation and protecting vulnerable audiences. This paper investigates susceptibility to influence within social networks, focusing on the differential effects of influence-driven versus spontaneous behaviors on user content adoption. Our analysis reveals that influence-driven adoption exhibits high homophily, indicating that individuals prone to influence often connect with similarly susceptible peers, thereby reinforcing peer influence dynamics, whereas spontaneous adoption shows significant but lower homophily. Additionally, we extend the Generalized Friendship Paradox to influence-driven behaviors, demonstrating that users' friends are generally more susceptible to influence than the users themselves, de facto establishing the notion of Susceptibility Paradox in online social influence. This pattern does not hold for spontaneous behaviors, where friends exhibit fewer spontaneous adoptions. We find that susceptibility to influence can be predicted using friends' susceptibility alone, while predicting spontaneous adoption requires additional features, such as user metadata. These findings highlight the complex interplay between user engagement and characteristics in spontaneous content adoption. Our results provide new insights into social influence mechanisms and offer implications for designing more effective moderation strategies to protect vulnerable audiences