X's Community Notes, a crowd-sourced fact-checking system, allows users to annotate potentially misleading posts.Notes rated as helpful by a diverse set of users are prominently displayed below the original post.While demonstrably effective at reducing misinformation's impact when notes are displayed, there is an opportunity for notes to appear on many more posts: for 91% of posts where at least one note is proposed, no notes ultimately achieve sufficient support from diverse users to be shown on the platform.This motivates the development of Supernotes: AI-generated notes that synthesize information from several existing community notes and are written to foster consensus among a diverse set of users.Our framework uses an LLM to generate many diverse Supernote candidates from existing proposed notes.These candidates are then evaluated by a novel scoring model, trained on millions of historical Community Notes ratings, selecting candidates that are most likely to be rated helpful by a diverse set of users.To test our framework, we ran a human subjects experiment in which we asked participants to compare the Supernotes generated by our framework to the best existing community notes for 100 sample posts.We found that participants rated the Supernotes as significantly more helpful, and when asked to choose between the two, preferred the Supernotes 75.2% of the time.Participants also rated the Supernotes more favorably than the best existing notes on quality, clarity, coverage, context, and argumentativeness.Finally, in a follow-up experiment, we asked participants to compare the Supernotes against LLMgenerated summaries and found that the participants rated the Supernotes significantly more helpful, demonstrating that both the LLM-based candidate generation and the consensus-driven scoring play crucial roles in creating notes that effectively build consensus among diverse users.

The goal of influence maximization is to select a set of seed users that will optimally diffuse information through a network. In this paper, we study how applying traditional influence maximization algorithms affects the balance between different audience categories (e.g., gender breakdown) who will eventually be exposed to a message. More specifically, we investigate how structural homophily (i.e., the tendency to connect to similar others) and influence diffusion homophily (i.e., the tendency to be influenced by similar others) affect the balance among the activated nodes. We find that even under mild levels of homophily, the balance among the exposed nodes is significantly worse than the balance among the overall population, resulting in a significant disadvantage for one group. To address this challenge, we propose an algorithm that jointly maximizes the influence and balance among nodes while still preserving the attractive theoretical guarantees of the traditional influence maximization algorithms. We run a series of experiments on multiple synthetic and four real-world datasets to demonstrate the effectiveness of the proposed algorithm in improving the balance between different categories of exposed nodes.

Social media platforms promise to enable rich and vibrant conversations online; however, their potential is often hindered by antisocial behaviors. In this paper, we study the relationship between structure and toxicity in conversations on Twitter. We collect 1.18M conversations (58.5M tweets, 4.4M users) prompted by tweets that are posted by or mention major news outlets over one year and candidates who ran in the 2018 US midterm elections over four months. We analyze the conversations at the individual, dyad, and group level. At the individual level, we find that toxicity is spread across many low to moderately toxic users. At the dyad level, we observe that toxic replies are more likely to come from users who do not have any social connection nor share many common friends with the poster. At the group level, we find that toxic conversations tend to have larger, wider, and deeper reply trees, but sparser follow graphs. To test the predictive power of the conversational structure, we consider two prediction tasks. In the first prediction task, we demonstrate that the structural features can be used to predict whether the conversation will become toxic as early as the first ten replies. In the second prediction task, we show that the structural characteristics of the conversation are also predictive of whether the next reply posted by a specific user will be toxic or not. We observe that the structural and linguistic characteristics of the conversations are complementary in both prediction tasks. Our findings inform the design of healthier social media platforms and demonstrate that models based on the structural characteristics of conversations can be used to detect early signs of toxicity and potentially steer conversations in a less toxic direction.

Homophily - our tendency to surround ourselves with others who share our perspectives and opinions about the world - is both a part of human nature and an organizing principle underpinning many of our digital social networks. However, when it comes to politics or culture, homophily can amplify tribal mindsets and produce "echo chambers" that degrade the quality, safety, and diversity of discourse online. While several studies have empirically proven this point, few have explored how making users aware of the extent and nature of their political echo chambers influences their subsequent beliefs and actions. In this paper, we introduce Social Mirror, a social network visualization tool that enables a sample of Twitter users to explore the politically-active parts of their social network. We use Social Mirror to recruit Twitter users with a prior history of political discourse to a randomized experiment where we evaluate the effects of different treatments on participants' i) beliefs about their network connections, ii) the political diversity of who they choose to follow, and iii) the political alignment of the URLs they choose to share. While we see no effects on average political alignment of shared URLs, we find that recommending accounts of the opposite political ideology to follow reduces participants» beliefs in the political homogeneity of their network connections but still enhances their connection diversity one week after treatment. Conversely, participants who enhance their belief in the political homogeneity of their Twitter connections have less diverse network connections 2-3 weeks after treatment. We explore the implications of these disconnects between beliefs and actions on future efforts to promote healthier exchanges in our digital public spheres.

Randomized experiments, or A/B tests, are the standard approach for evaluating the causal effects of new product features, i.e., treatments. The validity of these tests rests on the "stable unit treatment value assumption" (SUTVA), which implies that the treatment only affects the behavior of treated users, and does not affect the behavior of their connections. Violations of SUTVA, common in features that exhibit network effects, result in inaccurate estimates of the causal effect of treatment. In this paper, we leverage a new experimental design for testing whether SUTVA holds, without making any assumptions on how treatment effects may spill over between the treatment and the control group. To achieve this, we simultaneously run both a completely randomized and a cluster-based randomized experiment, and then we compare the difference of the resulting estimates. We present a statistical test for measuring the significance of this difference and offer theoretical bounds on the Type I error rate. We provide practical guidelines for implementing our methodology on large-scale experimentation platforms. Importantly, the proposed methodology can be applied to settings in which a network is not necessarily observed but, if available, can be used in the analysis. Finally, we deploy this design to LinkedIn's experimentation platform and apply it to two online experiments, highlighting the presence of network effects and bias in standard A/B testing approaches in a real-world setting.

Purchase logs collected in e-commerce platforms provide rich information about customer preferences. These logs can be leveraged to improve the quality of product recommendations by feeding them to machine-learned ranking models. However, a variety of deployment constraints limit the naive applicability of machine learning to this problem. First, the amount and the dimensionality of the data make in-memory learning simply not possible. Second, the drift of customers' preference over time require to retrain the ranking model regularly with freshly collected data. This limits the time that is available for training to prohibitively short intervals. Third, ranking in real-time is necessary whenever the query complexity prevents us from caching the predictions. This constraint requires to minimize prediction time (or equivalently maximize the data throughput), which in turn may prevent us from achieving the accuracy necessary in web-scale industrial applications. In this paper, we investigate how the practical challenges faced in this setting can be tackled via an online learning to rank approach. Sparse models will be the key to reduce prediction latency, whereas one-pass stochastic optimization will minimize the training time and restrict the memory footprint. Interestingly, and perhaps surprisingly, extensive experiments show that one-pass learning preserves most of the predictive performance. Additionally, we study a variety of online learning algorithms that enforce sparsity and provide insights to help the practitioner make an informed decision about which approach to pick. We report results on a massive purchase log dataset from the Amazon retail website, as well as on several benchmarks from the LETOR corpus.

Recommender systems suggest to users items that they might like (e.g., news articles, songs, movies) and, in doing so, they help users deal with information overload and enjoy a personalized experience. One of the main problems of these systems is the item cold-start, i.e., when a new item is introduced in the system and no past information is available, then no effective recommendations can be produced. The item cold-start is a very common problem in practice: modern online platforms have hundreds of new items published every day. To address this problem, we propose to learn Local Collective Embeddings: a matrix factorization that exploits items' properties and past user preferences while enforcing the manifold structure exhibited by the collective embeddings. We present a learning algorithm based on multiplicative update rules that are efficient and easy to implement. The experimental results on two item cold-start use cases: news recommendation and email recipient recommendation, demonstrate the effectiveness of this approach and show that it significantly outperforms six state-of-the-art methods for item cold-start.