Zoher Kachwala, Bao Tran Truong, Rasika Muralidharan, Haewoon Kwak, Jisun An, Filippo Menczer. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Psychiatric symptom identification on social media aims to infer fine-grained mental health symptoms from user-generated posts, allowing a detailed understanding of users' mental states. However, the construction of large-scale symptom-level datasets remains challenging due to the resource-intensive nature of expert labeling and the lack of standardized annotation guidelines, which in turn limits the generalizability of models to identify diverse symptom expressions from user-generated text. To address these issues, we propose SynSym, a synthetic data generation framework for constructing generalizable datasets for symptom identification. Leveraging large language models (LLMs), SynSym constructs high-quality training samples by (1) expanding each symptom into sub-concepts to enhance the diversity of generated expressions, (2) producing synthetic expressions that reflect psychiatric symptoms in diverse linguistic styles, and (3) composing realistic multi-symptom expressions, informed by clinical co-occurrence patterns. We validate SynSym on three benchmark datasets covering different styles of depressive symptom expression. Experimental results demonstrate that models trained solely on the synthetic data generated by SynSym perform comparably to those trained on real data, and benefit further from additional fine-tuning with real data. These findings underscore the potential of synthetic data as an alternative resource to real-world annotations in psychiatric symptom modeling, and SynSym serves as a practical framework for generating clinically relevant and realistic symptom expressions.

Quantifying identity fusion-the psychological merging of self with another entity or abstract target (e.g., a religious group, political party, ideology, value, brand, belief, etc.)-is vital for understanding a wide range of group-based human behaviors.We introduce the Cognitive Linguistic Identity Fusion Score (CLIFS), a novel metric that integrates cognitive linguistics with large language models (LLMs), which builds on implicit metaphor detection.Unlike traditional pictorial and verbal scales, which require controlled surveys or direct field contact, CLIFS delivers fully automated, scalable assessments while maintaining strong alignment with the established verbal measure.In benchmarks, CLIFS outperforms both existing automated approaches and human annotation.As a proof of concept, we apply CLIFS to violence risk assessment to demonstrate that it can improve violence risk assessment by more than 240%.Building on our identification of a new NLP task and early success, we underscore the need to develop larger, more diverse datasets that encompass additional fusion-target domains and cultural backgrounds to enhance generalizability and further advance this emerging area.CLIFS models and code are public at

As AI becomes more integral in our lives, the need for transparency and responsibility grows. While natural language explanations (NLEs) are vital for clarifying the reasoning behind AI decisions, evaluating them through human judgments is complex and resource-intensive due to subjectivity and the need for fine-grained ratings. This study explores the alignment between ChatGPT and human assessments across multiple scales (i.e., binary, ternary, and 7-Likert scale). We sample 300 data instances from three NLE datasets and collect 900 human annotations for both informativeness and clarity scores as the text quality measurement. We further conduct paired comparison experiments under different ranges of subjectivity scores, where the baseline comes from 8,346 human annotations. Our results show that ChatGPT aligns better with humans in more coarse-grained scales. Also, paired comparisons and dynamic prompting (i.e., providing semantically similar examples in the prompt) improve the alignment. This research advances our understanding of large language models’ capabilities to assess the text explanation quality in different configurations for responsible AI development.

Traffic prediction is one of the key elements to ensure the safety and convenience of citizens. Existing traffic prediction models primarily focus on deep learning architectures to capture spatial and temporal correlation. They often overlook the underlying nature of traffic. Specifically, the sensor networks in most traffic datasets do not accurately represent the actual road network exploited by vehicles, failing to provide insights into the traffic patterns in urban activities. To overcome these limitations, we propose an improved traffic prediction method based on graph convolution deep learning algorithms. We leverage human activity frequency data from National Household Travel Survey to enhance the inference capability of a causal relationship between activity and traffic patterns. Despite making minimal modifications to the conventional graph convolutional recurrent networks and graph convolutional transformer architectures, our approach achieves state-of-the-art performance without introducing excessive computational overhead.

Conspiracy theories are widely propagated on social media. Among various social media services, YouTube is one of the most influential sources of news and entertainment. This paper seeks to develop a dataset, YOUNICON, to enable researchers to perform conspiracy theory detection as well as classification of videos with conspiracy theories into different topics. YOUNICON is a dataset with a large collection of videos from suspicious channels that were identified to contain conspiracy theories in a previous study. Overall, YOUNICON will enable researchers to study trends in conspiracy theories and understand how individuals can interact with the conspiracy theory producing community or channel. Our data is available at: https://doi.org/10.5281/zenodo.7466262.

False information spreads on social media, and fact-checking is a potential countermeasure. However, there is a severe shortage of fact-checkers; an efficient way to scale fact-checking is desperately needed, especially in pandemics like COVID-19. In this study, we focus on spontaneous debunking by social media users, which has been missed in existing research despite its indicated usefulness for fact-checking and countering false information. Specifically, we characterize the tweets with false information, or fake tweets, that tend to be debunked and Twitter users who often debunk fake tweets. For this analysis, we create a comprehensive dataset of responses to fake tweets, annotate a subset of them, and build a classification model for detecting debunking behaviors. We find that most fake tweets are left undebunked, spontaneous debunking is slower than other forms of responses, and spontaneous debunking exhibits partisanship in political topics. These results provide actionable insights into utilizing spontaneous debunking to scale conventional fact-checking, thereby supplementing existing research from a new perspective.

This research characterises user engagement of approximately 3,000,000 news postings of 53 news outlets and 50,000,000 associated user comments during 8 months on 5 social media platforms (i.e. Facebook, Instagram, Twitter, YouTube, and Reddit). We investigate the effect of sentiments and topics on user engagement across four levels of user engagement expressions (i.e. views, likes, comments, cross-platform posting). We find that sentiments and topics differ by both news outlets and social media platforms, and both sentiments and topics by the four levels of user engagement expression. Finally, we predict a volume of four user engagement levels for given news content, with an 83% maximum average F1-score for the external posting of news articles from one platform to another using language and metadata features. Implications are that news outlets can benefit by developing a platform, sentiment and topic, and strategies to best achieve user engagement objectives.

On behalf of all the members of the organizing Committee, we are pleased to welcome all of you to IEEE/ACM ASONAM 2022.

Collecting good ideas is vital for organizations, especially companies, to retain their competitiveness. Social media is gathering attention as a place to extract ideas efficiently; however, the characteristics of ideas and the posters of ideas on social media are underexamined. Thus, this study aims to characterize spontaneous ideation contests among social media users by taking an event of Facebook’s name change to Meta as a case study. As a dataset, we comprehensively collect tweets containing new acronyms of Big Tech companies, which we treat as an "idea" in this work. In the analysis, we especially focus on the diversity of ideas, which would be the main reason for enlisting social media for idea generation. As the main results, we discovered that social media users offered a wider range of ideas than those in mainstream media. The follow-follower network of the users suggested that the users’ position on the network is related to the preferred ideas. Additionally, we discovered a link between the amount of user interaction on social media and the diversity of ideas. This study would promote the use of social media as a part of open innovation and co-creation processes in the industry.

Predicting the political bias and the factuality of reporting of entire news outlets are critical elements of media profiling, which is an understudied but an increasingly important research direction.The present level of proliferation of fake, biased, and propagandistic content online, has made it impossible to fact-check every single suspicious claim, either manually or automatically.Alternatively, we can profile entire news outlets and look for those that are likely to publish fake or biased content.This approach makes it possible to detect likely "fake news" the moment they are published, by simply checking the reliability of their source.From a practical perspective, political bias and factuality of reporting have a linguistic aspect but also a social context.Here, we study the impact of both, namely (i) what was written (i.e., what was published by the target medium, and how it describes itself on Twitter) vs. (ii) who read it (i.e., analyzing the readers of the target medium on Facebook, Twitter, and YouTube).We further study (iii) what was written about the target medium on Wikipedia.The evaluation results show that what was written matters most, and that putting all information sources together yields huge improvements over the current state-of-the-art.

As Internet users increasingly rely on social media sites to receive news, they are faced with a bewildering number of news media choices. For example, thousands of Facebook pages today are registered and categorized as some form of news media outlets. This situation boosted the so-called independent journalism, also known as alternative news media. Identifying and characterizing all the news pages that play an important role in news dissemination is key for understanding the news ecosystems of a country. In this work, we propose a graph-based semi-supervised method to measure the political bias of pages on most countries and show the political split of the alternative media, mainstream media, and public figures pages. We validate our method using the publicly available U.S. dataset and then apply it to Brazilian pages, where we found a larger number of right-wing pages in general, except for alternative news media.

Because word semantics can substantially change across communities and contexts, capturing domain-specific word semantics is an important challenge.Here, we propose SEMAXIS, a simple yet powerful framework to characterize word semantics using many semantic axes in wordvector spaces beyond sentiment.We demonstrate that SEMAXIS can capture nuanced semantic representations in multiple online communities.We also show that, when the sentiment axis is examined, SEMAXIS outperforms the state-of-theart approaches in building domain-specific sentiment lexicons.

In this research, we investigate if and how more photos than a single headshot can heighten the level of information provided by persona profiles. We conduct eye-tracking experiments and qualitative interviews with variations in the photos: a single headshot, a headshot and images of the persona in different contexts, and a headshot with pictures of different people representing key persona attributes. The results show that more contextual photos significantly improve the information end users derive from a persona profile; however, showing images of different people creates confusion and lowers the informativeness. Moreover, we discover that choice of pictures results in various interpretations of the persona that are biased by the end users' experiences and preconceptions. The results imply that persona creators should consider the design power of photos when creating persona profiles.

We present Automatic Persona Generation (APG), a methodology and system for quantitative persona generation using large amounts of online social media data. The system is operational, beta deployed with several client organizations in multiple industry verticals and ranging from small-to-medium sized enterprises to large multi-national corporations. Using a robust web framework and stable back-end database, APG is currently processing tens of millions of user interactions with thousands of online digital products on multiple social media platforms, such as Facebook and YouTube. APG identifies both distinct and impactful user segments and then creates persona descriptions by automatically adding pertinent features, such as names, photos, and personal attributes. We present the overall methodological approach, architecture development, and main system features. APG has a potential value for organizations distributing content via online platforms and is unique in its approach to persona generation. APG can be found online at https://persona.qcri.org.

To more effectively convey relevant information to end users of persona profiles, we conducted a user study consisting of 29 participants engaging with three persona layout treatments. We were interested in confusion engendered by the treatments on the participants, and conducted a within-subjects study in the actual work environment, using eye-tracking and talk-aloud data collection. We coded the verbal data into classes of informativeness and confusion and correlated it with fixations and durations on the Areas of Interests recorded by the eye-tracking device. We used various analysis techniques, including Mann-Whitney, regression, and Levenshtein distance, to investigate how confused users differed from non-confused users, what information of the personas caused confusion, and what were the predictors of confusion of end users of personas. We consolidate our various findings into a confusion ratio measure, which highlights in a succinct manner the most confusing elements of the personas. Findings show that inconsistencies among the informational elements of the persona generate the most confusion, especially with the elements of images and social media quotes. The research has implications for the design of personas and related information products, such as user profiling and customer segmentation.

We investigate the alignment of international attention of news media organizations within 193 countries with the expressed international interests of the public within those same countries from March 7, 2016 to April 14, 2017. We collect fourteen months of longitudinal data of online news from Unfiltered News and web search volume data from Google Trends and build a multiplex network of media attention and public attention in order to study its structural and dynamic properties. Structurally, the media attention and the public attention are both similar and different depending on the resolution of the analysis. For example, we find that 63.2% of the country-specific media and the public pay attention to different countries, but local attention flow patterns, which are measured by network motifs, are very similar. We also show that there are strong regional similarities with both media and public attention that is only disrupted by significantly major worldwide incidents (e.g., Brexit). Using Granger causality, we show that there are a substantial number of countries where media attention and public attention are dissimilar by topical interest. Our findings show that the media and public attention toward specific countries are often at odds, indicating that the public within these countries may be ignoring their country-specific news outlets and seeking other online sources to address their media needs and desires.

We develop a methodology to automate creating imaginary people, referred to as personas, by processing complex behavioral and demographic data of social media audiences. From a popular social media account containing more than 30 million interactions by viewers from 198 countries engaging with more than 4,200 online videos produced by a global media corporation, we demonstrate that our methodology has several novel accomplishments, including: (a) identifying distinct user behavioral segments based on the user content consumption patterns; (b) identifying impactful demographics groupings; and (c) creating rich persona descriptions by automatically adding pertinent attributes, such as names, photos, and personal characteristics. We validate our approach by implementing the methodology into an actual working system; we then evaluate it via quantitative methods by examining the accuracy of predicting content preference of personas, the stability of the personas over time, and the generalizability of the method via applying to two other datasets. Research findings show the approach can develop rich personas representing the behavior and demographics of real audiences using privacy-preserving aggregated online social media data from major online platforms. Results have implications for media companies and other organizations distributing content via online platforms.

In this work, we describe a methodology for leveraging large amounts of customer interaction data with online content from major social media platforms in order to isolate meaningful customer segments. The methodology is robust in that it can rapidly identify diverse customer segments using solely online behaviors and then associate these behavioral customer segments with the related distinct demographic segments, presenting a holistic picture of the customer base of an organization. We validate our methodology via the implementation of a working system that rapidly and in near real-time processes tens of millions of online customer interactions with content posted on major social media platforms in order to identify both the distinct behavioral segments and corresponding impactful demographic segments. We illustrate the functionality of the methodology with real data from a major online content provider with millions of online interactions from more than thirty countries. We further show one possible use for such information via the automatic generation of personas for an organization, which can be used for the formulation of marketing strategy, implementation of advertising plans, or development of products. The research results offer insights into competitive marketing and product preferences for the consumers of online digital content. We conclude with a discussion of areas for future work.

We propose a novel method for generating personas based on online user data for the increasingly common situation of content creators distributing products via online platforms. We use non-negative matrix factorization to identify user segments and develop personas by adding personality such as names and photos. Our approach can develop accurate personas representing real groups of people using online user data, versus relying on manually gathered data.

We built a multiplex media attention and disregard network (MADN) among 129 countries over 212 days. By characterizing the MADN from multiple levels, we found that it is formed primarily by skewed, hierarchical, and asymmetric relationships. Also, we found strong evidence that our news world is becoming a "global village." However, at the same time, unique attention blocks of the Middle East and North Africa (MENA) region, as well as Russia and its neighbors, still exist.

Using personas during interactive design has considerable potential for product and content development. Unfortunately, personas have typically been a fairly static technique. In this research, we validate an approach for creating personas in real time, based on analysis of actual social media data in an effort to automate the generation of personas. We validate that social media data can be implemented as an approach for automating generating personas in real time using actual YouTube social media data from a global media corporation that produces online digital content. Using the organization's YouTube channel, we collect demographic data, customer interactions, and topical interests, leveraging more than 188,000 subscriber profiles and more than 30 million user interactions. Then, we conduct statistical analysis on the social media data to determine whether the data could lead to the generation of valid personas based on statistically difference market segments. Findings show that customers can be segmented using product topics by gender and age based using social media data. However, our findings also show that the data is biased by the content created. The results offer insights into competitive marketing and product preferences for the consumers of the online digital content. Implications are that personas can be generated in real-time using social media data, instead of a time-consuming manual development process.

To bring their innovative ideas to market, those embarking in new ventures have to raise money, and, to do so, they have often resorted to banks and venture capitalists. Nowadays, they have an additional option: that of crowdfunding. The name refers to the idea that funds come from a network of people on the Internet who are passionate about supporting others' projects. One of the most popular crowdfunding sites is Kickstarter. In it, creators post descriptions of their projects and advertise them on social media sites (mainly Twitter), while investors look for projects to support. The most common reason for project failure is the inability of founders to connect with a sufficient number of investors, and that is mainly because hitherto there has not been any automatic way of matching creators and investors. We thus set out to propose different ways of recommending investors found on Twitter for specific Kickstarter projects. We do so by conducting hypothesis-driven analyses of pledging behavior and translate the corresponding findings into different recommendation strategies. The best strategy achieves, on average, 84% of accuracy in predicting a list of potential investors' Twitter accounts for any given project. Our findings also produced key insights about the whys and wherefores of investors deciding to support innovative efforts.

A social-based routing protocol for opportunistic networks considers the direct delivery as forwarding metrics. By ignoring the indirect delivery through intermediate nodes, it misses chances to find paths that are better in terms of delivery ratio and time. To overcome this limitation, we propose to incorporate transitivity, which considers the indirect delivery through intermediate nodes, as one of the forwarding metrics. We also found that some message forwards do not improve the delivery performance. To reduce the number of these useless forwards, the proposed scheme forwards messages to an encountered node when the increase of total utility value is greater than a threshold. Using a simulator with real world trace data sets, we compare the proposed scheme with the existing protocols, epidemic routing and SimBetTS. Compared with SimBetTS, the proposed scheme increases delivery ratio by 1.5 percent and decreases delay time by 2 percent while reducing overhead by 30 percent.