Opioid 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 the stigma faced by people using 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. However, these communities 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.

Online communities on Reddit are a popular choice among people with opioid use disorder (OUD) to seek information on drug use, withdrawal symptoms, and recovery. LLM-powered chatbots (e.g., ChatGPT) are widely being adopted as question-answer systems for health-related queries. However, such online health information seeking could potentially be hindered by myths and misinformation on OUD, misleading or causing genuine harm to people with OUD. In this work, we examine the prevalence of 5 OUD-related myths, on treatment models and patient characteristics, within human- (taken from Reddit) and LLM-generated responses to queries on OUD. We further explore the framing strategies used within responses (both human- and LLM-generated) promoting and countering the myths. We found that all 5 myths were more widespread within human-generated responses. In addition, myth-promoting responses adopted trustworthy and authoritative framings, compared to knowledge-imparting linguistic cues within those countering the myths. Our work offers recommendations to reduce online OUD misinformation.

Voice-based social media platforms that enable attendees to have real-time, ephemeral interactions with each other—such as X-Spaces, Discord, and Clubhouse—have seen considerable growth in recent years. While prior research on these spaces has predominantly focused on moderating harms, our work seeks to understand emergent practices employed by hosts to proactively shape their discussion space— focusing on the facilitation aspect of moderation duties. Drawing on facilitation strategies, we study these practices through three comprehensive studies using mixed-methods: survey of social-audio users, co-design interviews, and analyzing training sessions for hosts. Our findings reveal insights into the issues faced by hosts and attendees, current facilitation practices, opinions on technological solutions, and factors that could be responsible for some of the identified issues such as the available training for hosts. We found that hosts themselves are often significant sources of issues due to practices such as focusing more on self-promotion than facilitating discussions. In addition, host training sessions seem to encourage behaviors that contribute to the negative perception of hosts. We draw on outcomes from co-design interviews to guide the design of future tools to support hosts in facilitating social-audio spaces. Our findings provide insights that could help create a more positive experience for both hosts and attendees.

Recent years have seen a rise in social virtual reality (VR) platforms that allow people to interact in real-time through voice and gestures. The ephemeral nature of communication on these platforms can enable new forms of harmful behavior and new challenges for moderators. We performed virtual field research on three VR environments (AltspaceVR, Horizon Worlds, Rec Room). Based on observing 100 scheduled events, our analysis uncovered 13 distinct types of potentially harmful behaviors enabled by real-time voice, embodied interactions, and platform affordances. We witnessed potential harm at 45% of our observed events; only 24% of these incidents were addressed by moderators. To understand moderation practices, we conducted interviews with 11 moderators to investigate how they assess real-time interactions and how they operate within the current state of moderation tools. Our work sheds light on how moderation tools and practices must evolve to meet the new challenges of social VR.

Mai ElSherief, Caleb Ziems, David Muchlinski, Vaishnavi Anupindi, Jordyn Seybolt, Munmun De Choudhury, Diyi Yang. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. 2021.

Existing work on automated hate speech classification assumes that the dataset is fixed and the classes are pre-defined.However, the amount of data in social media increases every day, and the hot topics changes rapidly, requiring the classifiers to be able to continuously adapt to new data without forgetting the previously learned knowledge.This ability, referred to as lifelong learning, is crucial for the realword application of hate speech classifiers in social media.In this work, we propose lifelong learning of hate speech classification on social media.To alleviate catastrophic forgetting, we propose to use Variational Representation Learning (VRL) along with a memory module based on LB-SOINN (Load-Balancing Self-Organizing Incremental Neural Network).Experimentally, we show that combining variational representation learning and the LB-SOINN memory module achieves better performance than the commonly-used lifelong learning techniques.

Andrew Gaut, Tony Sun, Shirlyn Tang, Yuxin Huang, Jing Qian, Mai ElSherief, Jieyu Zhao, Diba Mirza, Elizabeth Belding, Kai-Wei Chang, William Yang Wang. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.

Tony Sun, Andrew Gaut, Shirlyn Tang, Yuxin Huang, Mai ElSherief, Jieyu Zhao, Diba Mirza, Elizabeth Belding, Kai-Wei Chang, William Yang Wang. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019.

Existing work on automated hate speech detection typically focuses on binary classification or on differentiating among a small set of categories.In this paper, we propose a novel method on a fine-grained hate speech classification task, which focuses on differentiating among 40 hate groups of 13 different hate group categories.We first explore the Conditional Variational Autoencoder (CVAE) (Larsen et al., 2016;Sohn et al., 2015) as a discriminative model and then extend it to a hierarchical architecture to utilize the additional hate category information for more accurate prediction.Experimentally, we show that incorporating the hate category information for training can significantly improve the classification performance and our proposed model outperforms commonly-used discriminative models.

In this paper we study social networks as an enabling technology for new applications and services leveraging, largely unutilized, opportunistic mobile encounters. More specifically, we quantify mobile user similarity and introduce a novel mathematical framework, grounded in information theory, to characterize fundamental limits and quantify the performance of sample knowledge sharing strategies. First, we introduce generalized, non-temporal and temporal profile structures, beyond geographic location, as a probability mass function. Second, we examine classic and information-theoretic similarity metrics using data in the public domain. A noticeable finding is that temporal metrics give lower similarity indices on the average (i.e., conservative) compared to non-temporal metrics, due to leveraging the wealth of information in the temporal dimension. Third, we introduce a novel mathematical framework that establishes fundamental limits for knowledge sharing among similar opportunistic users. Finally, we show numerical results quantifying the cumulative knowledge gain over time and its upper bound, the knowledge gain limit, using public smartphone data for the user behavior and mobility traces, in the case of fixed as well as mobile scenarios. The presented results provide valuable insights highlighting the key role of the introduced information-theoretic framework in motivating future research along this ripe research direction, studying diverse scenarios as well as novel knowledge sharing strategies.

No abstract available.