The behavior of Large Language Models (LLMs) as artificial social agents is largely unexplored, and we still lack extensive evidence of how these agents react to simple social stimuli. Testing the behavior of AI agents in classic Game Theory experiments provides a promising theoretical framework for evaluating the norms and values of these agents in archetypal social situations. In this work, we investigate the cooperative behavior of three LLMs (Llama2, Llama3, and GPT3.5) when playing the Iterated Prisoner's Dilemma against random adversaries displaying various levels of hostility. We introduce a systematic methodology to evaluate an LLM's comprehension of the game rules and its capability to parse historical gameplay logs for decision-making. We conducted simulations of games lasting for 100 rounds and analyzed the LLMs' decisions in terms of dimensions defined in the behavioral economics literature. We find that all models tend not to initiate defection but act cautiously, favoring cooperation over defection only when the opponent's defection rate is low. Overall, LLMs behave at least as cooperatively as the typical human player, although our results indicate some substantial differences among models. In particular, Llama2 and GPT3.5 are more cooperative than humans, and especially forgiving and non-retaliatory for opponent defection rates below 30%. More similar to humans, Llama3 exhibits consistently uncooperative and exploitative behavior unless the opponent always cooperates. Our systematic approach to the study of LLMs in game theoretical scenarios is a step towards using these simulations to inform practices of LLM auditing and alignment.

Social media platforms shape climate action discourse. Mapping these online conversations is essential for effective communication strategies. TikTok’s climate discussions are particularly relevant given its young, climate-concerned audience. In this work, we collect the first TikTok dataset on climate topics. We collected 590K videos from 14K creators along with their follower networks. By applying topic modeling to the video descriptions, we map the topics discussed on the platform on a climate taxonomy that we construct by consolidating existing categorizations. Results show TikTok creators primarily approach climate through the angle of lifestyle and dietary choices. By examining semantic connections between topics, we identified non-climate ”gateway” topics that could draw new audiences into climate discussions.

Social media play a key role in mobilizing collective action, holding the potential for studying the pathways that lead individuals to actively engage in addressing global challenges. However, quantitative research in this area has been limited by the absence of granular and large-scale ground truth about the level of participation in collective action among individual social media users. To address this limitation, we present a novel suite of text classifiers designed to identify expressions of participation in collective action from social media posts, in a topic-agnostic fashion. Grounded in the theoretical framework of social movement mobilization, our classification captures participation and categorizes it into four levels: recognizing collective issues, engaging in calls-to-action, expressing intention of action, and reporting active involvement. We constructed a labeled training dataset of Reddit comments through crowdsourcing, which we used to train BERT classifiers and fine-tune Llama3 models. Our findings show that smaller language models can reliably detect expressions of participation (weighted F1=0.71), and rival larger models in capturing nuanced levels of participation. By applying our methodology to Reddit, we illustrate its effectiveness as a robust tool for characterizing online communities in innovative ways compared to topic modeling, stance detection, and keyword-based methods. Our framework contributes to Computational Social Science research by providing a new source of reliable annotations useful for investigating the social dynamics of collective action.

Climate activism is crucial in stimulating collective societal and behavioral change towards sustainable practices through political pressure.Although multiple factors contribute to the participation in activism, their complex relationships and the scarcity of data on their interactions have restricted most prior research to studying them in isolation, thus preventing the development of a quantitative, causal understanding of why people approach activism.In this work, we develop a comprehensive causal model of how and why Reddit users engage with activist communities driving mass climate protests (mainly the 2019 Earth Strike, Fridays for Future, and Extinction Rebellion).Our framework, based on Stochastic Variational Inference applied to Bayesian Networks, learns the causal pathways over multiple time periods.Distinct from previous studies, our approach uses large-scale and fine-grained longitudinal data (2016 to 2022) to jointly model the roles of sociodemographic makeup, experience of extreme weather events, exposure to climate-related news, and social influence through online interactions.We find that among users interested in climate change, participation in online activist communities is indeed influenced by direct interactions with activists and largely by recent exposure to media coverage of climate protests.Among people aware of climate change, left-leaning people from lower socioeconomic backgrounds are particularly represented in online activist groups.Our findings offer empirical validation for theories of media influence and critical mass, and lay the foundations to inform interventions and future studies to foster public participation in collective action.

The increasing capability of Large Language Models to act as human-like social agents raises two important questions in the area of opinion dynamics. First, whether these agents can generate effective arguments that could be injected into the online discourse to steer the public opinion. Second, whether artificial agents can interact with each other to reproduce dynamics of persuasion typical of human social systems, opening up opportunities for studying synthetic social systems as faithful proxies for opinion dynamics in human populations. To address these questions, we designed a synthetic persuasion dialogue scenario on the topic of climate change, where a 'convincer' agent generates a persuasive argument for a 'skeptic' agent, who subsequently assesses whether the argument changed its internal opinion state. Different types of arguments were generated to incorporate different linguistic dimensions underpinning psycho-linguistic theories of opinion change. We then asked human judges to evaluate the persuasiveness of machine-generated arguments. Arguments that included factual knowledge, markers of trust, expressions of support, and conveyed status were deemed most effective according to both humans and agents, with humans reporting a marked preference for knowledge-based arguments. Our experimental framework lays the groundwork for future in-silico studies of opinion dynamics, and our findings suggest that artificial agents have the potential of playing an important role in collective processes of opinion formation in online social media.

Narratives can be powerful tools for inspiring action on pressing societal issues such as climate change. While social science theories offer frameworks for understanding the narratives that arise within collective movements, these are rarely applied to the vast data available from social media platforms, which play a significant role in shaping public opinion and mobilizing collective action. This gap in the empirical evaluation of online narratives limits our understanding of their relationship with public response. In this study, we focus on plant-based diets as a form of pro-environmental action and employ Natural Language Processing to operationalize a theoretical framework of moral narratives specific to the vegan movement. We apply this framework to narratives found in YouTube videos promoting environmental initiatives such as Veganuary, Meatless March, and No Meat May. Our analysis reveals that several narrative types, as defined by the theory, are empirically present in the data. To identify narratives with the potential to elicit positive public engagement, we used text processing to estimate the proportion of comments supporting collective action across narrative types. Video narratives advocating social fight, whether through protest or through efforts to convert others to the cause, are associated with a stronger sense of collective action in the respective comments. These narrative types also demonstrate increased semantic coherence and alignment between the message and public response, markers typically associated with successful collective action. Our work offers new insights into the complex factors that influence the emergence of collective action, thereby informing the development of effective communication strategies within social movements.

Decades of social science research identified ten fundamental dimensions that provide the conceptual building blocks to describe the nature of human relationships. Yet, it is not clear to what extent these concepts are expressed in everyday language and what role they have in shaping observable dynamics of social interactions. After annotating conversational text through crowdsourcing, we trained NLP tools to detect the presence of these types of interaction from conversations, and applied them to 160M messages written by geo-referenced Reddit users, 290k emails from the Enron corpus and 300k lines of dialogue from movie scripts. We show that social dimensions can be predicted purely from conversations with an AUC up to 0.98, and that the combination of the predicted dimensions suggests both the types of relationships people entertain (conflict vs. support) and the types of real-world communities (wealthy vs. deprived) they shape.

Ego-networks are fundamental structures in social graphs, yet the process of their evolution is still widely unexplored. In an online context, a key question is how link recommender systems may skew the growth of these networks, possibly restraining diversity. To shed light on this matter, we analyze the complete temporal evolution of 170M ego-networks extracted from Flickr and Tumblr, comparing links that are created spontaneously with those that have been algorithmically recommended. We find that the evolution of ego-networks is bursty, community-driven, and characterized by subsequent phases of explosive diameter increase, slight shrinking, and stabilization. Recommendations favor popular and well-connected nodes, limiting the diameter expansion. With a matching experiment aimed at detecting causal relationships from observational data, we find that the bias introduced by the recommendations fosters global diversity in the process of neighbor selection. Last, with two link prediction experiments, we show how insights from our analysis can be used to improve the effectiveness of social recommender systems.

With mobile shopping surging in popularity, people are spending ever more money on digital purchases through their mobile devices and phones. However, few large-scale studies of mobile shopping exist. In this paper we analyze a large data set consisting of more than 776M digital purchases made on Apple mobile devices that include songs, apps, and in-app purchases. We find that 61% of all the spending is on in-app purchases and that the top 1% of users are responsible for 59% of all the spending. These big spenders are more likely to be male and older, and less likely to be from the US. We study how they adopt and abandon individual app, and find that, after an initial phase of increased daily spending, users gradually lose interest: the delay between their purchases increases and the spending decreases with a sharp drop toward the end. Finally, we model the in-app purchasing behavior in multiple steps: 1) we model the time between purchases; 2) we train a classifier to predict whether the user will make a purchase from a new app or continue purchasing from the existing app; and 3) based on the outcome of the previous step, we attempt to predict the exact app, new or existing, from which the next purchase will come. The results yield new insights into spending habits in the mobile digital marketplace.

User participation in online communities is driven by the intertwinement of the social network structure with the crowd-generated content that flows along its links. These aspects are rarely explored jointly and at scale. By looking at how users generate and access pictures of varying beauty on Flickr, we investigate how the production of quality impacts the dynamics of online social systems. We develop a deep learning computer vision model to score images according to their aesthetic value and we validate its output through crowdsourcing. By applying it to over 15 B Flickr photos, we study for the first time how image beauty is distributed over a large-scale social system. Beautiful images are evenly distributed in the network, although only a small core of people get social recognition for them. To study the impact of exposure to quality on user engagement, we set up matching experiments aimed at detecting causality from observational data. Exposure to beauty is double-edged: following people who produce high-quality content increases one's probability of uploading better photos; however, an excessive imbalance between the quality generated by a user and the user's neighbors leads to a decline in engagement. Our analysis has practical implications for improving link recommender systems.

Sponsored search aims at retrieving the advertisements that in the one hand meet users' intent reflected in their search queries, and in the other hand attract user clicks to generate revenue. Advertisements are typically ranked based on their expected revenue that is computed as the product between their predicted probability of being clicked (i.e., namely clickability) and their advertiser provided bid. The relevance of an advertisement to a user query is implicitly captured by the predicted clickability of the advertisement, assuming that relevant advertisements are more likely to attract user clicks. However, this approach easily biases the ranking toward advertisements having rich click history. This may incorrectly lead to showing irrelevant advertisements whose clickability is not accurately predicted due to lack of click history. Another side effect consists of never giving a chance to new advertisements that may be highly relevant to be printed due to their lack of click history. To address this problem, we explicitly measure the relevance between an advertisement and a query without relying on the advertisement's click history, and present different ways of leveraging this relevance to improve user search experience without reducing search engine revenue. Specifically, we propose a machine learning approach that solely relies on text-based features to measure the relevance between an advertisement and a query. We discuss how the introduced relevance can be used in four important use cases: pre-filtering of irrelevant advertisements, recovering advertisements with little history, improving clickability prediction, and re-ranking of the advertisements on the final search result page. Offine experiments using large-scale query logs and online A/B tests demonstrate the superiority of the proposed click-oblivious relevance model and the important roles that relevance plays in sponsored search.

Consumer spending accounts for a large fraction of economic footprint of modern countries. Increasingly, consumer activity is moving to the web, where digital receipts of online purchases provide valuable data sources detailing consumer behavior. We consider such data extracted from emails and combined with with consumers' demographic information, which we use to characterize, model, and predict purchasing behavior. We analyze such behavior of consumers in different age and gender groups, and find interesting, actionable patterns that can be used to improve ad targeting systems. For example, we found that the amount of money spent on online purchases grows sharply with age, peaking in the late 30s, while shoppers from wealthy areas tend to purchase more expensive items and buy them more frequently. Furthermore, we look at the influence of social connections on purchasing habits, as well as at the temporal dynamics of online shopping where we discovered daily and weekly behavioral patterns. Finally, we build a model to predict when shoppers are most likely to make a purchase and how much will they spend, showing improvement over baseline approaches. The presented results paint a clear picture of a modern online shopper, and allow better understanding of consumer behavior that can help improve marketing efforts and make shopping more pleasant and efficient experience for online customers.

Community question answering (CQA) sites use a collaborative paradigm to satisfy complex information needs. Although the task of matching questions to their best answers has been tackled for more than a decade, the social question-answering practice is a complex process. The factors influencing the accuracy of question-answer matching are many and hard to disentangle. We approach the task from an application-oriented perspective, probing the space of several dimensions relevant to this problem: features, algorithms, and topics. We gather under a learning to rank framework the most extensive feature set used in literature to date, including 225 features from five different families. We test the power of such features in predicting the best answer to a question on the largest dataset from Yahoo Answers used for this task so far (40M answers) and provide a faceted analysis of the results along different topical areas and question types. We propose a novel family of distributional semantics measures that most of the time can seamlessly replace widely used linguistic similarity features, being more than one order of magnitude faster to compute and providing greater predictive power. The best feature set reaches an improvement between 11% and 26% in P@1 compared to recent well-established state-of-the-art methods.

The "Local Ranking Problem" (LRP) is related to the computation of a centrality-like rank on a local graph, where the scores of the nodes could significantly differ from the ones computed on the global graph. Previous work has studied LRP on the hyperlink graph but never on the BrowseGraph, namely a graph where nodes are webpages and edges are browsing transitions. Recently, this graph has received more and more attention in many different tasks such as ranking, prediction and recommendation. However, a web-server has only the browsing traffic performed on its pages (local BrowseGraph) and, as a consequence, the local computation can lead to estimation errors, which hinders the increasing number of applications in the state of the art. Also, although the divergence between the local and global ranks has been measured, the possibility of estimating such divergence using only local knowledge has been mainly overlooked. These aspects are of great interest for online service providers who want to: (i) gauge their ability to correctly assess the importance of their resources only based on their local knowledge, and (ii) take into account real user browsing fluxes that better capture the actual user interest than the static hyperlink network. We study the LRP problem on a BrowseGraph from a large news provider, considering as subgraphs the aggregations of browsing traces of users coming from different domains. We show that the distance between rankings can be accurately predicted based only on structural information of the local graph, being able to achieve an average rank correlation as high as 0.8.

Email is a ubiquitous communications tool in the workplace and plays an important role in social interactions. Previous studies of email were largely based on surveys and limited to relatively small populations of email users within organizations. In this paper, we report results of a large-scale study of more than 2 million users exchanging 16 billion emails over several months. We quantitatively characterize the replying behavior in conversations within pairs of users. In particular, we study the time it takes the user to reply to a received message and the length of the reply sent. We consider a variety of factors that affect the reply time and length, such as the stage of the conversation, user demographics, and use of portable devices. In addition, we study how increasing load affects emailing behavior. We find that as users receive more email messages in a day, they reply to a smaller fraction of them, using shorter replies. However, their responsiveness remains intact, and they may even reply to emails faster. Finally, we predict the time to reply, length of reply, and whether the reply ends a conversation. We demonstrate considerable improvement over the baseline in all three prediction tasks, showing the significant role that the factors that we uncover play, in determining replying behavior. We rank these factors based on their predictive power. Our findings have important implications for understanding human behavior and designing better email management applications for tasks like ranking unread emails.

Walkability has many health, environmental, and economic benefits. That is why web and mobile services have been offering ways of computing walkability scores of individual street segments. Those scores are generally computed from survey data and manual counting (of even trees). However, that is costly, owing to the high time, effort, and financial costs. To partly automate the computation of those scores, we explore the possibility of using the social media data of Flickr and Foursquare to automatically identify safe and walkable streets. We find that unsafe streets tend to be photographed during the day, while walkable streets are tagged with walkability-related keywords. These results open up practical opportunities (for, e.g., room booking services, urban route recommenders, and real-estate sites) and have theoretical implications for researchers who might resort to the use social media data to tackle previously unanswered questions in the area of walkability.

Online social networks and mash-up services create opportunities to connect different web services otherwise isolated. Specifically in the case of news, users are very much exposed to news articles while performing other activities, such as social networking or web searching. Browsing behavior aimed at the consumption of news, especially in relation to the visits coming from other domains, has been mainly overlooked in previous work. To address that, we build a BrowseGraph out of the collective browsing traces extracted from a large viewlog of Yahoo News (0.5B entries), and we define the ReferrerGraph as its subgraph induced by the sessions with the same referrer domain. The structural and temporal properties of the graph show that browsing behavior in news is highly dependent on the referrer URL of the session, in terms of type of content consumed and time of consumption. We build on this observation and propose a news recommender that addresses the cold-start problem: given a user landing on a page of the site for the first time, we aim to predict the page she will visit next. We compare 24 flavors of recommenders belonging to the families of content-based, popularity-based, and browsing-based models. We show that the browsing-based recommender that takes into account the referrer URL is the best performing, achieving a prediction accuracy of 48% in conditions of heavy data sparsity.

Finding experts in question answering platforms has important applications, such as question routing or identification of best answers. Addressing the problem of ranking users with respect to their expertise, we propose Competition-Based Expertise Networks (CBEN), a novel community expertise network structure based on the principle of competition among the answerers of a question. We evaluate our approach on a very large dataset from Yahoo! Answers using a variety of centrality measures. We show that it outperforms state-of-the-art network structures and, unlike previous methods, is able to consistly outperform simple metrics like best answer count. We also analyse question answering forums in Yahoo! Answers, and show that they can be characterised by factual or subjective information seeking behavior, social discussions and the conducting of polls or surveys. We find that the ability to identify experts greatly depends on the type of forum, which is directly reflected in the structural properties of the expertise networks.

Social groups play a crucial role in social media platforms because they form the basis for user participation and engagement. Groups are created explicitly by members of the community, but also form organically as members interact. Due to their importance, they have been studied widely (e.g., community detection, evolution, activity, etc.). One of the key questions for understanding how such groups evolve is whether there are different types of groups and how they differ. In Sociology, theories have been proposed to help explain how such groups form. In particular, the common identity and common bond theory states that people join groups based on identity (i.e., interest in the topics discussed) or bond attachment (i.e., social relationships). The theory has been applied qualitatively to small groups to classify them as either topical or social. We use the identity and bond theory to define a set of features to classify groups into those two categories. Using a dataset from Flickr, we extract user-defined groups and automatically-detected groups, obtained from a community detection algorithm. We discuss the process of manual labeling of groups into social or topical and present results of predicting the group label based on the defined features. We directly validate the predictions of the theory showing that the metrics are able to forecast the group type with high accuracy. In addition, we present a comparison between declared and detected groups along topicality and sociality dimensions.

Ranking of images is difficult because many factors determine their importance (e.g., popularity, quality, entertainment value, context, etc.). In social media platforms, ranking also depends on social interactions and on the visibility of the images both inside and outside those platforms. In this context, the application of standard ranking methods is not clearly understood, and neither are the subtleties associated with taking into account social interaction, internal, and external factors. In this paper, we use a large Flickr dataset and investigate these factors by performing an in-depth analysis of several ranking algorithms using both internal (i.e., within Flickr) and external (i.e., links from outside of Flickr) factors. We analyze rankings given by common metrics used in image retrieval (e.g., number of favorites), and compare them with metrics based on page views (e.g., time spent, number of views). In addition, we represent users' navigation by a graph and combine session models with some of these metrics, comparing with PageRank and BrowseRank. Our experiments show significant differences between the rankings, providing insights on the impact of social interactions, internal, and external factors in image ranking.

Social media have attracted considerable attention because their open-ended nature allows users to create lightweight semantic scaffolding to organize and share content. To date, the interplay of the social and topical components of social media has been only partially explored. Here, we study the presence of homophily in three systems that combine tagging social media with online social networks. We find a substantial level of topical similarity among users who are close to each other in the social network. We introduce a null model that preserves user activity while removing local correlations, allowing us to disentangle the actual local similarity between users from statistical effects due to the assortative mixing of user activity and centrality in the social network. This analysis suggests that users with similar interests are more likely to be friends, and therefore topical similarity measures among users based solely on their annotation metadata should be predictive of social links. We test this hypothesis on several datasets, confirming that social networks constructed from topical similarity capture actual friendship accurately. When combined with topological features, topical similarity achieves a link prediction accuracy of about 92%.

Categorization of web-search queries in semantically coherent topics is a crucial task to understand the interest trends of search engine users and, therefore, to provide more intelligent personalization services. Query clustering usually relies on lexical and clickthrough data, while the information originating from the user actions in submitting their queries is currently neglected. In particular, the intent that drives users to submit their requests is an important element for meaningful aggregation of queries. We propose a new intent-centric notion of topical query clusters and we define a query clustering technique that differs from existing algorithms in both methodology and nature of the resulting clusters. Our method extracts topics from the query log by merging missions, i.e., activity fragments that express a coherent user intent, on the basis of their topical affinity. Our approach works in a bottom-up way, without any a-priori knowledge of topical categorization, and produces good quality topics compared to state-of-the-art clustering techniques. It can also summarize topically-coherent missions that occur far away from each other, thus enabling a more compact user profiling on a topical basis. Furthermore, such a topical user profiling discriminates the stream of activity of a particular user from the activity of others, with a potential to predict future user search activity.