Generating social networks is essential for many applications, such as epidemic modeling and social simulations. The emergence of generative AI, especially large language models (LLMs), offers new possibilities for social network generation: LLMs can generate networks without additional training or need to define network parameters, and users can flexibly define individuals in the network using natural language. However, this potential raises two critical questions: 1) are the social networks generated by LLMs realistic, and 2) what are risks of bias, given the importance of demographics in forming social ties? To answer these questions, we develop three prompting methods for network generation and compare the generated networks to a suite of real social networks. We find that more realistic networks are generated with “local” methods, where the LLM constructs relations for one persona at a time, compared to “global” methods that construct the entire network at once. We also find that the generated networks match real networks on many characteristics, including density, clustering, connectivity, and degree distribution. However, we find that LLMs emphasize political homophily over all other types of homophily and significantly overestimate political homophily compared to real social networks.

Therapeutic peptides show promise in targeting previously undruggable binding sites, with recent advancements in deep generative models enabling full-atom peptide co-design for specific protein receptors.However, the critical role of molecular surfaces in proteinprotein interactions (PPIs) has been underexplored.To bridge this gap, we propose an omni-design peptides generation paradigm, called SurfFlow, a novel surface-based generative algorithm that enables comprehensive co-design of sequence, structure, and surface for peptides.SurfFlow employs a multi-modality conditional flow matching (CFM) architecture to learn distributions of surface geometries and biochemical properties, enhancing peptide binding accuracy.Evaluated on the comprehensive PepMerge benchmark, SurfFlow consistently outperforms full-atom baselines across all metrics.These results highlight the advantages of considering molecular surfaces in de novo peptide discovery and demonstrate the potential of integrating multiple protein modalities for more effective therapeutic peptide discovery.

Graph machine learning has led to a significant increase in the capabilities of models that learn on arbitrary graph-structured data and has been applied to molecules, social networks, recommendation systems, and transportation, among other domains. Data in multi-tabular relational databases can also be constructed as 'relational entity graphs' for Relational Deep Learning (RDL) - a new blueprint that enables end-to-end representation learning without traditional feature engineering. Compared to arbitrary graph-structured data, relational entity graphs have key properties: (i) their structure is defined by primary-foreign key relationships between entities in different tables(ii) the structural connectivity is a function of the relational schema defining a database, and (iii) the graph connectivity is temporal and heterogeneous in nature. In this paper, we provide a comprehensive review of RDL by first introducing the representation of relational databases as relational entity graphs, and then reviewing public benchmark datasets that have been used to develop and evaluate recent GNN-based RDL models. We discuss key challenges including large scale multi-table integration and the complexities of modeling temporal dynamics and heterogeneous data, while also surveying foundational neural network methods and recent architectural advances specialized for relational entity graphs. Finally, we explore opportunities to unify these distinct modeling challenges, highlighting how RDL converges multiple sub-fields in graph machine learning towards the design of foundation models that can transform the processing of relational data.

The web is the world's largest knowledge repository, yet as AI systems become increasingly integrated into our digital infrastructure, the ability to retrieve, reason, and collaborate effectively has become paramount. Large Language Models (LLMs) are evolving from passive responders to active knowledge agents that can retrieve complex information, validate hypotheses, and optimize interactions over multiple turns. In this talk, I will explore the frontiers of AI-driven knowledge retrieval and reasoning, drawing from recent research on knowledge graphs, semi-structured retrieval, adaptive tool use, and multi-turn AI collaboration. I will also discuss how agentic frameworks enable rigorous, automated hypothesis validation through sequential falsifications. Together, these advancements push beyond traditional search and QA systems, unlocking new capabilities for knowledge discovery, scientific research, and human-AI collaboration. Finally, I will highlight key challenges and opportunities in building AI systems that are not just accurate, but also interactive, explainable, and aligned with human needs.

With the recent progress in geometric deep learning, generative modeling, and the availability of large-scale biological datasets, molecular graph and geometry generative modeling have emerged as a highly promising direction for scientific discovery such as drug design. These generative methods enable efficient chemical space exploration and potential drug candidate generation. However, by representing molecules as 2D graphs or 3D geometries, there exist many both fundamental and challenging problems for modeling the distribution of these irregular and complex relational data. In this tutorial, we will introduce participants to the latest key developments in this field, covering important topics including 2D molecular graph generation, 3D molecular geometry generation, 2D graph to 3D geometry generation, and conditional 3D molecular geometry generation. We further include antibody generation, where we particularly consider large-size antibody molecules. For each topic, we will outline the underlying problem characteristics, summarize key challenges, present unified views of the representative approaches, and highlight future research direction and potential impacts. We anticipate this lecture-style tutorial would attract a broad audience of researchers and practitioners.

The era of intelligent systems and applications is here and AI has been disrupting science and industry. The speed of change has been breathtaking. However, building AI-based solutions is hard, training Machine Learning models and putting them in production takes highly-skilled teams months if not years. There is a real need to drastically simplify machine learning workflow and bring AI closer to users, and make it accessible to a wide range of audiences.

Learned embeddings for products are an important building block for web-scale e-commerce recommendation systems. At Pinterest, we build a single set of product embeddings called ItemSage to provide relevant recommendations in all shopping use cases including user, image and search based recommendations. This approach has led to significant improvements in engagement and conversion metrics, while reducing both infrastructure and maintenance cost. While most prior work focuses on building product embeddings from features coming from a single modality, we introduce a transformer-based architecture capable of aggregating information from both text and image modalities and show that it significantly outperforms single modality baselines. We also utilize multi-task learning to make ItemSage optimized for several engagement types, leading to a candidate generation system that is efficient for all of the engagement objectives of the end-to-end recommendation system. Extensive offline experiments are conducted to illustrate the effectiveness of our approach and results from online A/B experiments show substantial gains in key business metrics (up to +7% gross merchandise value/user and +11% click volume).

Embeddings, low-dimensional vector representation of objects, are fundamental in building modern machine learning systems. In industrial settings, there is usually an embedding team that trains an embedding model to solve intended tasks (e.g., product recommendation). The produced embeddings are then widely consumed by consumer teams to solve their unintended tasks (e.g., fraud detection). However, as the embedding model gets updated and retrained to improve performance on the intended task, the newly-generated embeddings are no longer compatible with the existing consumer models. This means that historical versions of the embeddings can never be retired or all consumer teams have to retrain their models to make them compatible with the latest version of the embeddings, both of which are extremely costly in practice.

Sequential models have become increasingly popular in powering personalized recommendation systems over the past several years. These approaches traditionally model a user's actions on a website as a sequence to predict the user's next action. While theoretically simplistic, these models are quite challenging to deploy in production, commonly requiring streaming infrastructure to reflect the latest user activity and potentially managing mutable data for encoding a user's hidden state. Here we introduce PinnerFormer, a user representation trained to predict a user's future long-term engagement using a sequential model of a user's recent actions. Unlike prior approaches, we adapt our modeling to a batch infrastructure via our new dense all-action loss, modeling long-term future actions instead of next action prediction. We show that by doing so, we significantly close the gap between batch user embeddings that are generated once a day and realtime user embeddings generated whenever a user takes an action. We describe our design decisions via extensive offline experimentation and ablations and validate the efficacy of our approach in A/B experiments showing substantial improvements in Pinterest's user retention and engagement when comparing PinnerFormer against our previous user representation. PinnerFormer is deployed in production as of Fall 2021.

Knowledge graphs (KGs) capture knowledge in the form of head--relation--tail triples and are a crucial component in many AI systems. There are two important reasoning tasks on KGs: (1) single-hop knowledge graph completion, which involves predicting individual links in the KG; and (2), multi-hop reasoning, where the goal is to predict which KG entities satisfy a given logical query. Embedding-based methods solve both tasks by first computing an embedding for each entity and relation, then using them to form predictions. However, existing scalable KG embedding frameworks only support single-hop knowledge graph completion and cannot be applied to the more challenging multi-hop reasoning task. Here we present Scalable Multi-hOp REasoning (SMORE), the first general framework for both single-hop and multi-hop reasoning in KGs. Using a single machine SMORE can perform multi-hop reasoning in Freebase KG (86M entities, 338M edges), which is 1,500x larger than previously considered KGs. The key to SMORE's runtime performance is a novel bidirectional rejection sampling that achieves a square root reduction of the complexity of online training data generation. Furthermore, SMORE exploits asynchronous scheduling, overlapping CPU-based data sampling, GPU-based embedding computation, and frequent CPU--GPU IO. SMORE increases throughput (i.e., training speed) over prior multi-hop KG frameworks by 2.2x with minimal GPU memory requirements (2GB for training 400-dim embeddings on 86M-node Freebase) and achieves near linear speed-up with the number of GPUs. Moreover, on the simpler single-hop knowledge graph completion task SMORE achieves comparable or even better runtime performance to state-of-the-art frameworks on both single GPU and multi-GPU settings.

Subsurface simulations use computational models to predict the flow of fluids (e.g., oil, water, gas) through porous media. These simulations are pivotal in industrial applications such as petroleum production, where fast and accurate models are needed for high-stake decision making, for example, for well placement optimization and field development planning. Classical finite difference numerical simulators require massive computational resources to model large-scale real-world reservoirs. Alternatively, streamline simulators and data-driven surrogate models are computationally more efficient by relying on approximate physics models, however they are insufficient to model complex reservoir dynamics at scale.

Graph Neural Networks (GNNs) have been successfully applied to many real-world static graphs. However, the success of static graphs has not fully translated to dynamic graphs due to the limitations in model design, evaluation settings, and training strategies. Concretely, existing dynamic GNNs do not incorporate state-of-the-art designs from static GNNs, which limits their performance. Current evaluation settings for dynamic GNNs do not fully reflect the evolving nature of dynamic graphs. Finally, commonly used training methods for dynamic GNNs are not scalable. Here we propose ROLAND, an effective graph representation learning framework for real-world dynamic graphs. At its core, the ROLAND framework can help researchers easily repurpose any static GNN to dynamic graphs. Our insight is to view the node embeddings at different GNN layers as hierarchical node states and then recurrently update them over time. We then introduce a live-update evaluation setting for dynamic graphs that mimics real-world use cases, where GNNs are making predictions and being updated on a rolling basis. Finally, we propose a scalable and efficient training approach for dynamic GNNs via incremental training and meta-learning. We conduct experiments over eight different dynamic graph datasets on future link prediction tasks. Models built using the ROLAND framework achieve on average 62.7% relative mean reciprocal rank (MRR) improvement over state-of-the-art baselines under the standard evaluation settings on three datasets. We find state-of-the-art baselines experience out-of-memory errors for larger datasets, while ROLAND can easily scale to dynamic graphs with 56 million edges. After re-implementing these baselines using the ROLAND training strategy, ROLAND models still achieve on average 15.5% relative MRR improvement over the baselines.

Graph Convolutional Networks (GCN) can efficiently integrate graph structure and node features to learn high-quality node embeddings. At Pinterest, we have developed and deployed PinSage, a data-efficient GCN that learns pin embeddings from the Pin-Board graph. Pinterest relies heavily on PinSage which in turn only leverages the Pin-Board graph. However, there exist several entities at Pinterest and heterogeneous interactions among these entities. These diverse entities and interactions provide important signal for recommendations and modeling. In this work, we show that training deep learning models on graphs that captures these diverse interactions can result in learning higher-quality pin embeddings than training PinSage on only the Pin-Board graph. However, building a large-scale heterogeneous graph engine that can process the entire Pinterest size data has not yet been done. In this work, we present a clever and effective solution where we break the heterogeneous graph into multiple disjoint bipartite graphs and then develop novel data-efficient MultiBiSage model that combines the signals from them. MultiBiSage can capture the graph structure of multiple bipartite graphs to learn high-quality pin embeddings. The benefit of our approach is that individual bipartite graphs can be processed with minimal changes to Pinterest's current infrastructure, while being able to combine information from all the graphs while achieving high performance. We train MultiBiSage on six bipartite graphs including our Pin-Board graph and show that it significantly outperforms the deployed latest version of PinSage on multiple user engagement metrics. We also perform experiments on two public datasets to show that MultiBiSage is generalizable and can be applied to datasets outside of Pinterest.

Label Propagation Algorithm (LPA) and Graph Convolutional Neural Networks (GCN) are both message passing algorithms on graphs. Both solve the task of node classification, but LPA propagates node label information across the edges of the graph, while GCN propagates and transforms node feature information. However, while conceptually similar, theoretical relationship between LPA and GCN has not yet been systematically investigated. Moreover, it is unclear how LPA and GCN can be combined under a unified framework to improve the performance. Here we study the relationship between LPA and GCN in terms of feature/label influence , in which we characterize how much the initial feature/label of one node influences the final feature/label of another node in GCN/LPA. Based on our theoretical analysis, we propose an end-to-end model that combines GCN and LPA. In our unified model, edge weights are learnable, and the LPA serves as regularization to assist the GCN in learning proper edge weights that lead to improved performance. Our model can also be seen as learning the weights of edges based on node labels, which is more direct and efficient than existing feature-based attention models or topology-based diffusion models. In a number of experiments for semi-supervised node classification and knowledge-graph-aware recommendation, our model shows superiority over state-of-the-art baselines.

Knowledge graph completion aims to predict missing relations between entities in a knowledge graph. In this work, we propose a relational message passing method for knowledge graph completion. Different from existing embedding-based methods, relational message passing only considers edge features (i.e., relation types) without entity IDs in the knowledge graph, and passes relational messages among edges iteratively to aggregate neighborhood information. Specifically, two kinds of neighborhood topology are modeled for a given entity pair under the relational message passing framework: (1) Relational context, which captures the relation types of edges adjacent to the given entity pair; (2) Relational paths, which characterize the relative position between the given two entities in the knowledge graph. The two message passing modules are combined together for relation prediction. Experimental results on knowledge graph benchmarks as well as our newly proposed dataset show that, our method PathCon outperforms state-of-the-art knowledge graph completion methods by a large margin. PathCon is also shown applicable to inductive settings where entities are not seen in training stage, and it is able to provide interpretable explanations for the predicted results. The code and all datasets are available at https://github.com/hwwang55/PathCon.

Mobility restrictions have been a primary intervention for controlling the spread of COVID-19, but they also place a significant economic burden on individuals and businesses. To balance these competing demands, policymakers need analytical tools to assess the costs and benefits of different mobility reduction measures. In this paper, we present our work motivated by our interactions with the Virginia Department of Health on a decision-support tool that utilizes large-scale data and epidemiological modeling to quantify the impact of changes in mobility on infection rates. Our model captures the spread of COVID-19 by using a fine-grained, dynamic mobility network that encodes the hourly movements of people from neighborhoods to individual places, with over 3 billion hourly edges. By perturbing the mobility network, we can simulate a wide variety of reopening plans and forecast their impact in terms of new infections and the loss in visits per sector. To deploy this model in practice, we built a robust computational infrastructure to support running millions of model realizations, and we worked with policymakers to develop an interactive dashboard that communicates our model's predictions for thousands of potential policies.

Abusive behavior in online retail websites and communities threatens the experience of regular community members. Such behavior often takes place within a complex, dynamic, and large-scale network of users interacting with items. Detecting abuse is challenging due to the scarcity of labeled abuse instances and complexity of combining temporal and network patterns while operating at a massive scale. Previous approaches to dynamic graph modeling either do not scale, do not effectively generalize from a few labeled instances, or compromise performance for scalability. Here we present BiDyn, a general method to detect abusive behavior in dynamic bipartite networks at scale, while generalizing from limited training labels. BiDyn develops an efficient hybrid RNN-GNN architecture trained via a novel stacked ensemble training scheme. We also propose a novel pre-training framework for dynamic graphs that helps to achieve superior performance at scale. Our approach outperforms recent large-scale dynamic graph baselines in an abuse classification task by up to 14% AUROC while requiring 10x less memory per training batch in both open and proprietary datasets.

Edge streams are commonly used to capture interactions in dynamic networks, such as email, social, or computer networks. The problem of detecting anomalies or rare events in edge streams has a wide range of applications. However, it presents many challenges due to lack of labels, a highly dynamic nature of interactions, and the entanglement of temporal and structural changes in the network. Current methods are limited in their ability to address the above challenges and to efficiently process a large number of interactions. Here, we propose F-FADE, a new approach for detection of anomalies in edge streams, which uses a novel frequency-factorization technique to efficiently model the time-evolving distributions of frequencies of interactions between node-pairs. The anomalies are then determined based on the likelihood of the observed frequency of each incoming interaction. F-FADE is able to handle in an online streaming setting a broad variety of anomalies with temporal and structural changes, while requiring only constant memory. Our experiments on one synthetic and six real-world dynamic networks show that F-FADE achieves state of the art performance and may detect anomalies that previous methods are unable to find.

Dynamic social interaction networks are an important abstraction to model time-stamped social interactions such as eye contact, speaking and listening between people. These networks typically contain informative while subtle patterns that reflect people’s social characters and relationship, and therefore attract the attentions of a lot of social scientists and computer scientists. Previous approaches on extracting those patterns primarily rely on sophisticated expert knowledge of psychology and social science, and the obtained features are often overly task-specific. More generic models based on representation learning of dynamic networks may be applied, but the unique properties of social interactions cause severe model mismatch and degenerate the quality of the obtained representations. Here we fill this gap by proposing a novel framework, termed TEmporal network-DIffusion Convolutional networks (TEDIC), for generic representation learning on dynamic social interaction networks. We make TEDIC a good fit by designing two components: 1) Adopt diffusion of node attributes over a combination of the original network and its complement to capture long-hop interactive patterns embedded in the behaviors of people making or avoiding contact; 2) Leverage temporal convolution networks with hierarchical set-pooling operation to flexibly extract patterns from different-length interactions scattered over a long time span. The design also endows TEDIC with certain self-explaining power. We evaluate TEDIC over five real datasets for four different social character prediction tasks including deception detection, dominance identification, nervousness detection and community detection. TEDIC not only consistently outperforms previous SOTA’s, but also provides two important pieces of social insight. In addition, it exhibits favorable societal characteristics by remaining unbiased to people from different regions. Our project website is: http://snap.stanford.edu/tedic/.

Graph Neural Networks (GNNs) are based on repeated aggregations of information from nodes' neighbors in a graph. However, because nodes share many neighbors, a naive implementation leads to repeated and inefficient aggregations and represents significant computational overhead. Here we propose Hierarchically Aggregated computation Graphs(HAGs), a new GNN representation technique that explicitly avoids redundancy by managing intermediate aggregation results hierarchically and eliminates repeated computations and unnecessary data transfers in GNN training and inference. HAGs perform the same computations and give the same models/accuracy as traditional GNNs, but in a much shorter time dueto optimized computations. To identify redundant computations,we introduce an accurate cost function and use a novel search algorithm to find optimized HAGs. Experiments show that the HAG representation significantly outperforms the standard GNN by increasing the end-to-end training throughput by up to 2.8× and reducing the aggregations and data transfers in GNN training byup to 6.3× and 5.6×, with only 0.1% memory overhead. Overall,our results represent an important advancement in speeding-up and scaling-up GNNs without any loss in model predictive performance.

Latent user representations are widely adopted in the tech industry for powering personalized recommender systems. Most prior work infers a single high dimensional embedding to represent a user, which is a good starting point but falls short in delivering a full understanding of the user's interests. In this work, we introduce PinnerSage, an end-to-end recommender system that represents each user via multi-modal embeddings and leverages this rich representation of users to provides high quality personalized recommendations. PinnerSage achieves this by clustering users' actions into conceptually coherent clusters with the help of a hierarchical clustering method (Ward) and summarizes the clusters via representative pins (Medoids) for efficiency and interpretability. PinnerSage is deployed in production at Pinterest and we outline the several design decisions that makes it run seamlessly at a very large scale. We conduct several offline and online A/B experiments to show that our method significantly outperforms single embedding methods.

Graph convolutional networks (GCNs) are a powerful class of graph neural networks. Trained in a semi-supervised end-to-end fashion, GCNs can learn to integrate node features and graph structures to generate high-quality embeddings that can be used for various downstream tasks like search and recommendation. However, existing GCNs mostly work on homogeneous graphs and consider a single embedding for each node, which do not sufficiently model the multi-facet nature and complex interaction of nodes in real-world networks. Here, we present a contextualized GCN engine by modeling the multipartite networks of target nodes and their intermediatecontext nodes that specify the contexts of their interactions. Towards the neighborhood aggregation process, we devise a contextual masking operation at the feature level and a contextual attention mechanism at the node level to achieve interaction contextualization by treating neighboring target nodes based on intermediate context nodes. Consequently, we compute multiple embeddings for target nodes that capture their diverse facets and different interactions during graph convolution, which is useful for fine-grained downstream applications. To enable efficient web-scale training, we build a parallel random walk engine to pre-sample contextualized neighbors, and a Hadoop2-based data provider pipeline to pre-join training data, dynamically reduce multi-GPU training time, and avoid high memory cost. Extensive experiments on the bipartite Pinterest graph and tripartite OAG graph corroborate the advantage of the proposed system.

Curated taxonomies enhance the performance of machine-learning systems via high-quality structured knowledge. However, manually curating a large and rapidly-evolving taxonomy is infeasible. In this work, we propose Arborist, an approach to automatically expand textual taxonomies by predicting the parents of new taxonomy nodes. Unlike previous work, Arborist handles the more challenging scenario of taxonomies with heterogeneous edge semantics that are unobserved. Arborist learns latent representations of the edge semantics along with embeddings of the taxonomy nodes to measure taxonomic relatedness between node pairs. Arborist is then trained by optimizing a large-margin ranking loss with a dynamic margin function. We propose a principled formulation of the margin function, which theoretically guarantees that Arborist minimizes an upper-bound on the shortest-path distance between the predicted parents and actual parents in the taxonomy. Via extensive evaluation on a curated taxonomy at Pinterest and several public datasets, we demonstrate that Arborist outperforms the state-of-the-art, achieving up to 59% in mean reciprocal rank and 83% in recall at 15. We also explore the ability of Arborist to infer nodes’ taxonomic-roles, without explicit supervision on this task.

Modeling sequential interactions between users and items/products is crucial in domains such as e-commerce, social networking, and education. Representation learning presents an attractive opportunity to model the dynamic evolution of users and items, where each user/item can be embedded in a Euclidean space and its evolution can be modeled by an embedding trajectory in this space. However, existing dynamic embedding methods generate embeddings only when users take actions and do not explicitly model the future trajectory of the user/item in the embedding space. Here we propose JODIE, a coupled recurrent neural network model that learns the embedding trajectories of users and items. JODIE employs two recurrent neural networks to update the embedding of a user and an item at every interaction. Crucially, JODIE also models the future embedding trajectory of a user/item. To this end, it introduces a novel projection operator that learns to estimate the embedding of the user at any time in the future. These estimated embeddings are then used to predict future user-item interactions. To make the method scalable, we develop a t-Batch algorithm that creates time-consistent batches and leads to 9x faster training. We conduct six experiments to validate JODIE on two prediction tasks---future interaction prediction and state change prediction---using four real-world datasets. We show that JODIE outperforms six state-of-the-art algorithms in these tasks by at least 20% in predicting future interactions and 12% in state change prediction.

Knowledge graphs capture structured information and relations between a set of entities or items. As such knowledge graphs represent an attractive source of information that could help improve recommender systems. However, existing approaches in this domain rely on manual feature engineering and do not allow for an end-to-end training. Here we propose Knowledge-aware Graph Neural Networks with Label Smoothness regularization (KGNN-LS) to provide better recommendations. Conceptually, our approach computes user-specific item embeddings by first applying a trainable function that identifies important knowledge graph relationships for a given user. This way we transform the knowledge graph into a user-specific weighted graph and then apply a graph neural network to compute personalized item embeddings. To provide better inductive bias, we rely on label smoothness assumption, which posits that adjacent items in the knowledge graph are likely to have similar user relevance labels/scores. Label smoothness provides regularization over the edge weights and we prove that it is equivalent to a label propagation scheme on a graph. We also develop an efficient implementation that shows strong scalability with respect to the knowledge graph size. Experiments on four datasets show that our method outperforms state of the art baselines. KGNN-LS also achieves strong performance in cold-start scenarios where user-item interactions are sparse.

The phenomenon of edge clustering in real-world networks is a fundamental property underlying many ideas and techniques in network science. Clustering is typically quantified by the clustering coefficient, which measures the fraction of pairs of neighbors of a given center node that are connected. However, many common explanations of edge clustering attribute the triadic closure to a head node instead of the center node of a length-2 path; for example, a friend of my friend is also my friend. While such explanations are common in network analysis, there is no measurement for edge clustering that can be attributed to the head node. Here we develop local closure coefficients as a metric quantifying head-node-based edge clustering. We define the local closure coefficient as the fraction of length-2 paths emanating from the head node that induce a triangle. This subtle difference in definition leads to remarkably different properties from traditional clustering coefficients. We analyze correlations with node degree, connect the closure coefficient to community detection, and show that closure coefficients as a feature can improve link prediction.

that goal. Despite its crucial role in how a number of prominent self-tracking apps function, there has been relatively little investigation of the goal-setting and achievement aspects of self-tracking apps. Here we explore this issue, investigating a particular goal setting and achievement process that is extensive, recorded, and crucial for both the app and its users' success: weight loss goals in MyFitnessPal. We present a large-scale study of 1.4 million users and weight loss goals, allowing for an unprecedented detailed view of how people set and achieve their goals. We find that, even for difficult long-term goals, behavior within the first 7 days predicts those who ultimately achieve their goals, that is, those who lose at least as much weight as they set out to, and those who do not. For instance, high amounts of early weight loss, which some researchers have classified as unsustainable, leads to higher goal achievement rates. We also show that early food intake, self-monitoring motivation, and attitude towards the goal are important factors. We then show that we can use our findings to predict goal achievement with an accuracy of 79% ROC AUC just 7 days after a goal is set. Finally, we discuss how our findings could inform steps to improve goal achievement in self-tracking apps.

Predicting pregnancy has been a fundamental problem in women's health for more than 50 years. Previous datasets have been collected via carefully curated medical studies, but the recent growth of women's health tracking mobile apps offers potential for reaching a much broader population. However, the feasibility of predicting pregnancy from mobile health tracking data is unclear. Here we develop four models - a logistic regression model, and 3 LSTM models - to predict a woman's probability of becoming pregnant using data from a women's health tracking app, Clue by BioWink GmbH. Evaluating our models on a dataset of 79 million logs from 65,276 women with ground truth pregnancy test data, we show that our predicted pregnancy probabilities meaningfully stratify women: women in the top 10% of predicted probabilities have a 89% chance of becoming pregnant over 6 menstrual cycles, as compared to a 27% chance for women in the bottom 10%. We develop a technique for extracting interpretable time trends from our deep learning models, and show these trends are consistent with previous fertility research. Our findings illustrate the potential that women's health tracking data offers for predicting pregnancy on a broader population; we conclude by discussing the steps needed to fulfill this potential.

Recommender systems that can learn from cross-session data to dynamically predict the next item a user will choose are crucial for online platforms. However, existing approaches often use out-of-the-box sequence models which are limited by speed and memory consumption, are often infeasible for production environments, and usually do not incorporate cross-session information, which is crucial for effective recommendations. Here we propose Hierarchical Temporal Convolutional Networks (HierTCN), a hierarchical deep learning architecture that makes dynamic recommendations based on users' sequential multi-session interactions with items. HierTCN is designed for web-scale systems with billions of items and hundreds of millions of users. It consists of two levels of models: The high-level model uses Recurrent Neural Networks (RNN) to aggregate users' evolving long-term interests across different sessions, while the low-level model is implemented with Temporal Convolutional Networks (TCN), utilizing both the long-term interests and the short-term interactions within sessions to predict the next interaction. We conduct extensive experiments on a public XING dataset and a large-scale Pinterest dataset that contains 6 million users with 1.6 billion interactions. We show that HierTCN is 2.5x faster than RNN-based models and uses 90% less data memory compared to TCN-based models. We further develop an effective data caching scheme and a queue-based mini-batch generator, enabling our model to be trained within 24 hours on a single GPU. Our model consistently outperforms state-of-the-art dynamic recommendation methods, with up to 18% improvement in recall and 10% in mean reciprocal rank.

Nodes residing in different parts of a graph can have similar structural roles within their local network topology. The identification of such roles provides key insight into the organization of networks and can be used for a variety of machine learning tasks. However, learning structural representations of nodes is a challenging problem, and it has typically involved manually specifying and tailoring topological features for each node. In this paper, we develop GraphWave, a method that represents each node's network neighborhood via a low-dimensional embedding by leveraging heat wavelet diffusion patterns. Instead of training on hand-selected features, GraphWave learns these embeddings in an unsupervised way. We mathematically prove that nodes with similar network neighborhoods will have similar GraphWave embeddings even though these nodes may reside in very different parts of the network, and our method scales linearly with the number of edges. Experiments in a variety of different settings demonstrate GraphWave's real-world potential for capturing structural roles in networks, and our approach outperforms existing state-of-the-art baselines in every experiment, by as much as 137%.

Recent advancements in deep neural networks for graph-structured data have led to state-of-the-art performance on recommender system benchmarks. However, making these methods practical and scalable to web-scale recommendation tasks with billions of items and hundreds of millions of users remains an unsolved challenge. Here we describe a large-scale deep recommendation engine that we developed and deployed at Pinterest. We develop a data-efficient Graph Convolutional Network (GCN) algorithm, which combines efficient random walks and graph convolutions to generate embeddings of nodes (i.e., items) that incorporate both graph structure as well as node feature information. Compared to prior GCN approaches, we develop a novel method based on highly efficient random walks to structure the convolutions and design a novel training strategy that relies on harder-and-harder training examples to improve robustness and convergence of the model. We also develop an efficient MapReduce model inference algorithm to generate embeddings using a trained model. Overall, we can train on and embed graphs that are four orders of magnitude larger than typical GCN implementations. We show how GCN embeddings can be used to make high-quality recommendations in various settings at Pinterest, which has a massive underlying graph with 3 billion nodes representing pins and boards, and 17 billion edges. According to offline metrics, user studies, as well as A/B tests, our approach generates higher-quality recommendations than comparable deep learning based systems. To our knowledge, this is by far the largest application of deep graph embeddings to date and paves the way for a new generation of web-scale recommender systems based on graph convolutional architectures.

Misinformation and misbehavior mining on the web(MIS2) workshop is held in Los Angeles, California, USA on February 9, 2018, and co-located with the 11th ACM International Conference on Web Search and Data Mining(WSDM 2018). Web is a dynamic ecosystem that enables malicious users to create and spread deceptive information to a wide audience in a matter of minutes. These malicious actors work on a wide variety of platforms, such as social media, e-commerce, and more. The main object of MIS2 is to discuss new and upcoming research on modeling, discovery, detection, and mitigation methods of misbehavior and misinformation on the web. MIS2 is an interdisciplinary venue for leading researchers and practitioners from the areas of data mining, social network analysis, cybersecurity, communications, human-computer interaction, and natural language processing. The topics addressed in MIS2 are extremely timely and the research presented by refereed papers and invited keynote speakers will participants a full dose of emerging research.

Mobile health applications that track activities, such as exercise, sleep, and diet, are becoming widely used. While these activity tracking applications have the potential to improve our health, user engagement and retention are critical factors for their success. However, long-term user engagement patterns in real-world activity tracking applications are not yet well understood. Here we study user engagement patterns within a mobile physical activity tracking application consisting of 115 million logged activities taken by over a million users over 31 months. Specifically, we show that over 75% of users return and re-engage with the application after prolonged periods of inactivity, no matter the duration of the inactivity. We find a surprising result that the re-engagement usage patterns resemble those of the start of the initial engagement period, rather than being a simple continuation of the end of the initial engagement period. This evidence points to a conceptual model of multiple lives of user engagement, extending the prevalent single life view of user activity. We demonstrate that these multiple lives occur because the users have a variety of different primary intents or goals for using the app. These primary intents are associated with how long each life lasts and how likely the user is to re-engage for a new life. We find evidence for users being more likely to stop using the app once they achieved their primary intent or goal (e.g., weight loss). However, these users might return once their original intent resurfaces (e.g., wanting to lose newly gained weight). We discuss implications of the multiple life paradigm and propose a novel prediction task of predicting the number of lives of a user. Based on insights developed in this work, including a marker of improved primary intent performance, our prediction models achieve 71% ROC AUC. Overall, our research has implications for modeling user re-engagement in health activity tracking applications and has consequences for how notifications, recommendations as well as gamification can be used to increase engagement.

User experience in modern content discovery applications critically depends on high-quality personalized recommendations. However, building systems that provide such recommendations presents a major challenge due to a massive pool of items, a large number of users, and requirements for recommendations to be responsive to user actions and generated on demand in real-time. Here we present Pixie, a scalable graph-based real-time recommender system that we developed and deployed at Pinterest. Given a set of user-specific pins as a query, Pixie selects in real-time from billions of possible pins those that are most related to the query. To generate recommendations, we develop Pixie Random Walk algorithm that utilizes the Pinterest object graph of 3 billion nodes and 17 billion edges. Experiments show that recommendations provided by Pixie lead up to 50% higher user engagement when compared to the previous Hadoop-based production system. Furthermore, we develop a graph pruning strategy at that leads to an additional 58% improvement in recommendations. Last, we discuss system aspects of Pixie, where a single server executes 1,200 recommendation requests per second with 60 millisecond latency. Today, systems backed by Pixie contribute to more than 80% of all user engagement on Pinterest.

Users organize themselves into communities on web platforms. These communities can interact with one another, often leading to conflicts and toxic interactions. However, little is known about the mechanisms of interactions between communities and how they impact users.

Mobile health applications, including those that track activities such as exercise, sleep, and diet, are becoming widely used. Accurately predicting human actions in the real world is essential for targeted recommendations that could improve our health and for personalization of these applications. However, making such predictions is extremely difficult due to the complexities of human behavior, which consists of a large number of potential actions that vary over time, depend on each other, and are periodic. Previous work has not jointly modeled these dynamics and has largely focused on item consumption patterns instead of broader types of behaviors such as eating, commuting or exercising. In this work, we develop a novel statistical model, called TIPAS, for Time-varying, Interdependent, and Periodic Action Sequences. Our approach is based on personalized, multivariate temporal point processes that model time-varying action propensities through a mixture of Gaussian intensities. Our model captures short-term and long-term periodic interdependencies between actions through Hawkes process-based self-excitations. We evaluate our approach on two activity logging datasets comprising 12 million real-world actions (e.g., eating, sleep, and exercise) taken by 20 thousand users over 17 months. We demonstrate that our approach allows us to make successful predictions of future user actions and their timing. Specifically, TIPAS improves predictions of actions, and their timing, over existing methods across multiple datasets by up to 156%, and up to 37%, respectively. Performance improvements are particularly large for relatively rare and periodic actions such as walking and biking, improving over baselines by up to 256%. This demonstrates that explicit modeling of dependencies and periodicities in real-world behavior enables successful predictions of future actions, with implications for modeling human behavior, app personalization, and targeting of health interventions.

Cycles are fundamental to human health and behavior. Examples include mood cycles, circadian rhythms, and the menstrual cycle. However, modeling cycles in time series data is challenging because in most cases the cycles are not labeled or directly observed and need to be inferred from multidimensional measurements taken over time. Here, we present Cyclic Hidden Markov Models (CyHMMs) for detecting and modeling cycles in a collection of multidimensional heterogeneous time series data. In contrast to previous cycle modeling methods, CyHMMs deal with a number of challenges encountered in modeling real-world cycles: they can model multivariate data with both discrete and continuous dimensions; they explicitly model and are robust to missing data; and they can share information across individuals to accommodate variation both within and between individual time series. Experiments on synthetic and real-world health-tracking data demonstrate that CyHMMs infer cycle lengths more accurately than existing methods, with 58% lower error on simulated data and 63% lower error on real-world data compared to the best-performing baseline. CyHMMs can also perform functions which baselines cannot: they can model the progression of individual features/symptoms over the course of the cycle, identify the most variable features, and cluster individual time series into groups with distinct characteristics. Applying CyHMMs to two real-world health-tracking datasets -- of human menstrual cycle symptoms and physical activity tracking data -- yields important insights including which symptoms to expect at each point during the cycle. We also find that people fall into several groups with distinct cycle patterns, and that these groups differ along dimensions not provided to the model. For example, by modeling missing data in the menstrual cycles dataset, we are able to discover a medically relevant group of birth control users even though information on birth control is not given to the model.

Machine learning on graphs is an important and ubiquitous task with applications ranging from drug design to friendship recommendation in social networks. The primary challenge in this domain is finding a way to represent, or encode, graph structure so that it can be easily exploited by machine learning models. However, traditionally machine learning approaches rely on user-defined heuristics to extract features encoding structural information about a graph. In this talk I will discuss methods that automatically learn to encode graph structure into low-dimensional embeddings, using techniques based on deep learning and nonlinear dimensionality reduction. I will provide a conceptual review of key advancements in this area of representation learning on graphs, including random-walk based algorithms, and graph convolutional networks.

Information items draw collective attention across a heterogeneous social system, leading to great disparities of popularity. Unveiling underlying diffusion processes is very challenging, since a social system consists of time-evolving subgroups interacting and exerting disproportionate influences on an individual item's popularity. In this study, we propose the Affinity Poisson Process model (APP) which models popularity dynamics, by incorporating (1) affinities between subgroups, (2) heterogeneous preferential attachment, and (3) subgroup-level time decay. As a case study, we apply our proposed model to scholarly publications in computer science. Our model outperforms the state of the art approach in predicting citation volumes of individual papers. More importantly, the proposed model enables us to uncover popularity dynamics driven by intra- and inter-subgroup interactions, which has been neglected in prior work. We expect that our model can afford interpretable insights on the attention economy in terms of affinity and aging effect.

Many important problems can be modeled as a system of interconnected entities, where each entity is recording time-dependent observations or measurements. In order to spot trends, detect anomalies, and interpret the temporal dynamics of such data, it is essential to understand the relationships between the different entities and how these relationships evolve over time. In this paper, we introduce the time-varying graphical lasso (TVGL), a method of inferring time-varying networks from raw time series data. We cast the problem in terms of estimating a sparse time-varying inverse covariance matrix, which reveals a dynamic network of interdependencies between the entities. Since dynamic network inference is a computationally expensive task, we derive a scalable message-passing algorithm based on the Alternating Direction Method of Multipliers (ADMM) to solve this problem in an efficient way. We also discuss several extensions, including a streaming algorithm to update the model and incorporate new observations in real time. Finally, we evaluate our TVGL algorithm on both real and synthetic datasets, obtaining interpretable results and outperforming state-of-the-art baselines in terms of both accuracy and scalability.

Subsequence clustering of multivariate time series is a useful tool for discovering repeated patterns in temporal data. Once these patterns have been discovered, seemingly complicated datasets can be interpreted as a temporal sequence of only a small number of states, or clusters. For example, raw sensor data from a fitness-tracking application can be expressed as a timeline of a select few actions (i.e., walking, sitting, running). However, discovering these patterns is challenging because it requires simultaneous segmentation and clustering of the time series. Furthermore, interpreting the resulting clusters is difficult, especially when the data is high-dimensional. Here we propose a new method of model-based clustering, which we call Toeplitz Inverse Covariance-based Clustering (TICC). Each cluster in the TICC method is defined by a correlation network, or Markov random field (MRF), characterizing the interdependencies between different observations in a typical subsequence of that cluster. Based on this graphical representation, TICC simultaneously segments and clusters the time series data. We solve the TICC problem through alternating minimization, using a variation of the expectation maximization (EM) algorithm. We derive closed-form solutions to efficiently solve the two resulting subproblems in a scalable way, through dynamic programming and the alternating direction method of multipliers (ADMM), respectively. We validate our approach by comparing TICC to several state-of-the-art baselines in a series of synthetic experiments, and we then demonstrate on an automobile sensor dataset how TICC can be used to learn interpretable clusters in real-world scenarios.

Evaluating whether machines improve on human performance is one of the central questions of machine learning. However, there are many domains where the data is selectively labeled, in the sense that the observed outcomes are themselves a consequence of the existing choices of the human decision-makers. For instance, in the context of judicial bail decisions, we observe the outcome of whether a defendant fails to return for their court appearance only if the human judge decides to release the defendant on bail. This selective labeling makes it harder to evaluate predictive models as the instances for which outcomes are observed do not represent a random sample of the population. Here we propose a novel framework for evaluating the performance of predictive models on selectively labeled data. We develop an approach called contraction which allows us to compare the performance of predictive models and human decision-makers without resorting to counterfactual inference. Our methodology harnesses the heterogeneity of human decision-makers and facilitates effective evaluation of predictive models even in the presence of unmeasured confounders (unobservables) which influence both human decisions and the resulting outcomes. Experimental results on real world datasets spanning diverse domains such as health care, insurance, and criminal justice demonstrate the utility of our evaluation metric in comparing human decisions and machine predictions.

Local graph clustering methods aim to find a cluster of nodes by exploring a small region of the graph. These methods are attractive because they enable targeted clustering around a given seed node and are faster than traditional global graph clustering methods because their runtime does not depend on the size of the input graph. However, current local graph partitioning methods are not designed to account for the higher-order structures crucial to the network, nor can they effectively handle directed networks. Here we introduce a new class of local graph clustering methods that address these issues by incorporating higher-order network information captured by small subgraphs, also called network motifs. We develop the Motif-based Approximate Personalized PageRank (MAPPR) algorithm that finds clusters containing a seed node with minimal \emph{motif conductance}, a generalization of the conductance metric for network motifs. We generalize existing theory to prove the fast running time (independent of the size of the graph) and obtain theoretical guarantees on the cluster quality (in terms of motif conductance). We also develop a theory of node neighborhoods for finding sets that have small motif conductance, and apply these results to the case of finding good seed nodes to use as input to the MAPPR algorithm. Experimental validation on community detection tasks in both synthetic and real-world networks, shows that our new framework MAPPR outperforms the current edge-based personalized PageRank methodology.

Many of today's most widely used computing applications utilize social networking features and allow users to connect, follow each other, share content, and comment on others' posts. However, despite the widespread adoption of these features, there is little understanding of the consequences that social networking has on user retention, engagement, and online as well as offline behavior. Here, we study how social networks influence user behavior in a physical activity tracking application. We analyze 791 million online and offline actions of 6 million users over the course of 5 years, and show that social networking leads to a significant increase in users' online as well as offline activities. Specifically, we establish a causal effect of how social networks influence user behavior. We show that the creation of new social connections increases user online in-application activity by 30%, user retention by 17%, and user offline real-world physical activity by 7% (about 400 steps per day). By exploiting a natural experiment we distinguish the effect of social influence of new social connections from the simultaneous increase in user's motivation to use the app and take more steps. We show that social influence accounts for 55% of the observed changes in user behavior, while the remaining 45% can be explained by the user's increased motivation to use the app. Further, we show that subsequent, individual edge formations in the social network lead to significant increases in daily steps. These effects diminish with each additional edge and vary based on edge attributes and user demographics. Finally, we utilize these insights to develop a model that accurately predicts which users will be most influenced by the creation of new social network connections.

Networks are a fundamental tool for modeling complex systems in a variety of domains including social and communication networks as well as biology and neuroscience. The counts of small subgraph patterns in networks, called network motifs, are crucial to understanding the structure and function of these systems. However, the role of network motifs for temporal networks, which contain many timestamped links between nodes, is not well understood.

In online discussion communities, users can interact and share information and opinions on a wide variety of topics. However, some users may create multiple identities, or sockpuppets, and engage in undesired behavior by deceiving others or manipulating discussions. In this work, we study sockpuppetry across nine discussion communities, and show that sockpuppets differ from ordinary users in terms of their posting behavior, linguistic traits, as well as social network structure. Sockpuppets tend to start fewer discussions, write shorter posts, use more personal pronouns such as ``I'', and have more clustered ego-networks. Further, pairs of sockpuppets controlled by the same individual are more likely to interact on the same discussion at the same time than pairs of ordinary users. Our analysis suggests a taxonomy of deceptive behavior in discussion communities. Pairs of sockpuppets can vary in their deceptiveness, i.e., whether they pretend to be different users, or their supportiveness, i.e., if they support arguments of other sockpuppets controlled by the same user. We apply these findings to a series of prediction tasks, notably, to identify whether a pair of accounts belongs to the same underlying user or not. Altogether, this work presents a data-driven view of deception in online discussion communities and paves the way towards the automatic detection of sockpuppets.

Wikipedia is one of the most popular sites on the Web, with millions of users relying on it to satisfy a broad range of information needs every day. Although it is crucial to understand what exactly these needs are in order to be able to meet them, little is currently known about why users visit Wikipedia. The goal of this paper is to fill this gap by combining a survey of Wikipedia readers with a log-based analysis of user activity. Based on an initial series of user surveys, we build a taxonomy of Wikipedia use cases along several dimensions, capturing users' motivations to visit Wikipedia, the depth of knowledge they are seeking, and their knowledge of the topic of interest prior to visiting Wikipedia. Then, we quantify the prevalence of these use cases via a large-scale user survey conducted on live Wikipedia with almost 30,000 responses. Our analyses highlight the variety of factors driving users to Wikipedia, such as current events, media coverage of a topic, personal curiosity, work or school assignments, or boredom. Finally, we match survey responses to the respondents' digital traces in Wikipedia's server logs, enabling the discovery of behavioral patterns associated with specific use cases. For instance, we observe long and fast-paced page sequences across topics for users who are bored or exploring randomly, whereas those using Wikipedia for work or school spend more time on individual articles focused on topics such as science. Our findings advance our understanding of reader motivations and behavior on Wikipedia and can have implications for developers aiming to improve Wikipedia's user experience, editors striving to cater to their readers' needs, third-party services (such as search engines) providing access to Wikipedia content, and researchers aiming to build tools such as recommendation engines.

Prediction tasks over nodes and edges in networks require careful effort in engineering features used by learning algorithms. Recent research in the broader field of representation learning has led to significant progress in automating prediction by learning the features themselves. However, present feature learning approaches are not expressive enough to capture the diversity of connectivity patterns observed in networks. Here we propose node2vec, an algorithmic framework for learning continuous feature representations for nodes in networks. In node2vec, we learn a mapping of nodes to a low-dimensional space of features that maximizes the likelihood of preserving network neighborhoods of nodes. We define a flexible notion of a node's network neighborhood and design a biased random walk procedure, which efficiently explores diverse neighborhoods. Our algorithm generalizes prior work which is based on rigid notions of network neighborhoods, and we argue that the added flexibility in exploring neighborhoods is the key to learning richer representations.

One of the most important obstacles to deploying predictive models is the fact that humans do not understand and trust them. Knowing which variables are important in a model's prediction and how they are combined can be very powerful in helping people understand and trust automatic decision making systems. Here we propose interpretable decision sets, a framework for building predictive models that are highly accurate, yet also highly interpretable. Decision sets are sets of independent if-then rules. Because each rule can be applied independently, decision sets are simple, concise, and easily interpretable. We formalize decision set learning through an objective function that simultaneously optimizes accuracy and interpretability of the rules. In particular, our approach learns short, accurate, and non-overlapping rules that cover the whole feature space and pay attention to small but important classes. Moreover, we prove that our objective is a non-monotone submodular function, which we efficiently optimize to find a near-optimal set of rules. Experiments show that interpretable decision sets are as accurate at classification as state-of-the-art machine learning techniques. They are also three times smaller on average than rule-based models learned by other methods. Finally, results of a user study show that people are able to answer multiple-choice questions about the decision boundaries of interpretable decision sets and write descriptions of classes based on them faster and more accurately than with other rule-based models that were designed for interpretability. Overall, our framework provides a new approach to interpretable machine learning that balances accuracy, interpretability, and computational efficiency.

Online e-commerce applications are becoming a primary vehicle for people to find, compare, and ultimately purchase products. One of the fundamental questions that arises in e-commerce is to characterize, understand, and model user long-term purchasing intent, which is important as it allows for personalized and context relevant e-commerce services.

Good websites should be easy to navigate via hyperlinks, yet maintaining a high-quality link structure is difficult. Identifying pairs of pages that should be linked may be hard for human editors, especially if the site is large and changes frequently. Further, given a set of useful link candidates, the task of incorporating them into the site can be expensive, since it typically involves humans editing pages. In the light of these challenges, it is desirable to develop data-driven methods for automating the link placement task. Here we develop an approach for automatically finding useful hyperlinks to add to a website. We show that passively collected server logs, beyond telling us which existing links are useful, also contain implicit signals indicating which nonexistent links would be useful if they were to be introduced. We leverage these signals to model the future usefulness of yet nonexistent links. Based on our model, we define the problem of link placement under budget constraints and propose an efficient algorithm for solving it. We demonstrate the effectiveness of our approach by evaluating it on Wikipedia, a large website for which we have access to both server logs (used for finding useful new links) and the complete revision history (containing a ground truth of new links). As our method is based exclusively on standard server logs, it may also be applied to any other website, as we show with the example of the biomedical research site Simtk.

Cascades of information-sharing are a primary mechanism by which content reaches its audience on social media, and an active line of research has studied how such cascades, which form as content is reshared from person to person, develop and subside. In this paper, we perform a large-scale analysis of cascades on Facebook over significantly longer time scales, and find that a more complex picture emerges, in which many large cascades recur, exhibiting multiple bursts of popularity with periods of quiescence in between. We characterize recurrence by measuring the time elapsed between bursts, their overlap and proximity in the social network, and the diversity in the demographics of individuals participating in each peak. We discover that content virality, as revealed by its initial popularity, is a main driver of recurrence, with the availability of multiple copies of that content helping to spark new bursts. Still, beyond a certain popularity of content, the rate of recurrence drops as cascades start exhausting the population of interested individuals. We reproduce these observed patterns in a simple model of content recurrence simulated on a real social network. Using only characteristics of a cascade's initial burst, we demonstrate strong performance in predicting whether it will recur in the future.

Wikipedia is a major source of information for many people. However, false information on Wikipedia raises concerns about its credibility. One way in which false information may be presented on Wikipedia is in the form of hoax articles, i.e., articles containing fabricated facts about nonexistent entities or events. In this paper we study false information on Wikipedia by focusing on the hoax articles that have been created throughout its history. We make several contributions. First, we assess the real-world impact of hoax articles by measuring how long they survive before being debunked, how many pageviews they receive, and how heavily they are referred to by documents on the Web. We find that, while most hoaxes are detected quickly and have little impact on Wikipedia, a small number of hoaxes survive long and are well cited across the Web. Second, we characterize the nature of successful hoaxes by comparing them to legitimate articles and to failed hoaxes that were discovered shortly after being created. We find characteristic differences in terms of article structure and content, embeddedness into the rest of Wikipedia, and features of the editor who created the hoax. Third, we successfully apply our findings to address a series of classification tasks, most notably to determine whether a given article is a hoax. And finally, we describe and evaluate a task involving humans distinguishing hoaxes from non-hoaxes. We find that humans are not particularly good at the task and that our automated classifier outperforms them by a big margin.

The different Wikipedia language editions vary dramatically in how comprehensive they are. As a result, most language editions contain only a small fraction of the sum of information that exists across all Wikipedias. In this paper, we present an approach to filling gaps in article coverage across different Wikipedia editions. Our main contribution is an end-to-end system for recommending articles for creation that exist in one language but are missing in another. The system involves identifying missing articles, ranking the missing articles according to their importance, and recommending important missing articles to editors based on their interests. We empirically validate our models in a controlled experiment involving 12,000 French Wikipedia editors. We find that personalizing recommendations increases editor engagement by a factor of two. Moreover, recommending articles increases their chance of being created by a factor of 3.2. Finally, articles created as a result of our recommendations are of comparable quality to organically created articles. Overall, our system leads to more engaged editors and faster growth of Wikipedia with no effect on its quality.

Large networks are becoming a widely used abstraction for studying complex systems in a broad set of disciplines, ranging from social network analysis to molecular biology and neuroscience. Despite an increasing need to analyze and manipulate large networks, only a limited number of tools are available for this task. Here, we describe Stanford Network Analysis Platform (SNAP), a general-purpose, high-performance system that provides easy to use, high-level operations for analysis and manipulation of large networks. We present SNAP functionality, describe its implementational details, and give performance benchmarks. SNAP has been developed for single big-memory machines and it balances the trade-off between maximum performance, compact in-memory graph representation, and the ability to handle dynamic graphs where nodes and edges are being added or removed over time. SNAP can process massive networks with hundreds of millions of nodes and billions of edges. SNAP offers over 140 different graph algorithms that can efficiently manipulate large graphs, calculate structural properties, generate regular and random graphs, and handle attributes and meta-data on nodes and edges. Besides being able to handle large graphs, an additional strength of SNAP is that networks and their attributes are fully dynamic, they can be modified during the computation at low cost. SNAP is provided as an open source library in C++ as well as a module in Python. We also describe the Stanford Large Network Dataset, a set of social and information real-world networks and datasets, which we make publicly available. The collection is a complementary resource to our SNAP software and is widely used for development and benchmarking of graph analytics algorithms.

No abstract available.

Convex optimization is an essential tool for modern data analysis, as it provides a framework to formulate and solve many problems in machine learning and data mining. However, general convex optimization solvers do not scale well, and scalable solvers are often specialized to only work on a narrow class of problems. Therefore, there is a need for simple, scalable algorithms that can solve many common optimization problems. In this paper, we introduce the network lasso, a generalization of the group lasso to a network setting that allows for simultaneous clustering and optimization on graphs. We develop an algorithm based on the Alternating Direction Method of Multipliers (ADMM) to solve this problem in a distributed and scalable manner, which allows for guaranteed global convergence even on large graphs. We also examine a non-convex extension of this approach. We then demonstrate that many types of problems can be expressed in our framework. We focus on three in particular --- binary classification, predicting housing prices, and event detection in time series data --- comparing the network lasso to baseline approaches and showing that it is both a fast and accurate method of solving large optimization problems.

To design a useful recommender system, it is important to understand how products relate to each other. For example, while a user is browsing mobile phones, it might make sense to recommend other phones, but once they buy a phone, we might instead want to recommend batteries, cases, or chargers. In economics, these two types of recommendations are referred to as substitutes and complements: substitutes are products that can be purchased instead of each other, while complements are products that can be purchased in addition to each other. Such relationships are essential as they help us to identify items that are relevant to a user's search.

Social networking websites allow users to create and share content. Big information cascades of post resharing can form as users of these sites reshare others' posts with their friends and followers. One of the central challenges in understanding such cascading behaviors is in forecasting information outbreaks, where a single post becomes widely popular by being reshared by many users. In this paper, we focus on predicting the final number of reshares of a given post. We build on the theory of self-exciting point processes to develop a statistical model that allows us to make accurate predictions. Our model requires no training or expensive feature engineering. It results in a simple and efficiently computable formula that allows us to answer questions, in real-time, such as: Given a post's resharing history so far, what is our current estimate of its final number of reshares? Is the post resharing cascade past the initial stage of explosive growth? And, which posts will be the most reshared in the future?

Spectral graph theory-based methods represent an important class of tools for studying the structure of networks. Spectral methods are based on a first-order Markov chain derived from a random walk on the graph and thus they cannot take advantage of important higher-order network substructures such as triangles, cycles, and feed-forward loops. Here we propose a Tensor Spectral Clustering (TSC) algorithm that allows for modeling higher-order network structures in a graph partitioning framework. Our TSC algorithm allows the user to specify which higher-order network structures (cycles, feed-forward loops, etc.) should be preserved by the network clustering. Higher-order network structures of interest are represented using a tensor, which we then partition by developing a multilinear spectral method. Our framework can be applied to discovering layered flows in networks as well as graph anomaly detection, which we illustrate on synthetic networks. In directed networks, a higher-order structure of particular interest is the directed 3-cycle, which captures feedback loops in networks. We demonstrate that our TSC algorithm produces large partitions that cut fewer directed 3-cycles than standard spectral clustering algorithms.

Several situations that we come across in our daily lives involve some form of evaluation: a process where an evaluator chooses a correct label for a given item. Examples of such situations include a crowd-worker labeling an image or a student answering a multiple-choice question. Gaining insights into human evaluations is important for determining the quality of individual evaluators as well as identifying true labels of items. Here, we generalize the question of estimating the quality of individual evaluators, extending it to obtain diagnostic insights into how various evaluators label different kinds of items. We propose a series of increasingly powerful hierarchical Bayesian models which infer latent groups of evaluators and items with the goal of obtaining insights into the underlying evaluation process. We apply our framework to a wide range of real-world domains, and demonstrate that our approach can accurately predict evaluator decisions, diagnose types of mistakes evaluators tend to make, and infer true labels of items.

We present Ringo, a system for analysis of large graphs. Graphs provide a way to represent and analyze systems of interacting objects (people, proteins, webpages) with edges between the objects denoting interactions (friendships, physical interactions, links). Mining graphs provides valuable insights about individual objects as well as the relationships among them. In building Ringo, we take advantage of the fact that machines with large memory and many cores are widely available and also relatively affordable. This allows us to build an easy-to-use interactive high-performance graph analytics system. Graphs also need to be built from input data, which often resides in the form of relational tables. Thus, Ringo provides rich functionality for manipulating raw input data tables into various kinds of graphs. Furthermore, Ringo also provides over 200 graph analytics functions that can then be applied to constructed graphs. We show that a single big-memory machine provides a very attractive platform for performing analytics on all but the largest graphs as it offers excellent performance and ease of use as compared to alternative approaches. With Ringo, we also demonstrate how to integrate graph analytics with an iterative process of trial-and-error data exploration and rapid experimentation, common in data mining workloads.

Recommender systems are an integral part of how we experience the Web today and they have become so ubiquitous that we do not even notice them anymore. However, today's recommender systems mostly treat items they recommend as black boxes and primarily focus on extracting correlations and co-counts from user behavior data. In this talk I argue that next generation recommender systems will require deep understanding of items being recommended as well as modeling the relationships between those items. I will present examples how auxiliary data about items (descriptions, reviews, product specifications) can be used to improve recommendations.

Hyperlinks are an essential feature of the World Wide Web. They are especially important for online encyclopedias such as Wikipedia: an article can often only be understood in the context of related articles, and hyperlinks make it easy to explore this context. But important links are often missing, and several methods have been proposed to alleviate this problem by learning a linking model based on the structure of the existing links. Here we propose a novel approach to identifying missing links in Wikipedia. We build on the fact that the ultimate purpose of Wikipedia links is to aid navigation. Rather than merely suggesting new links that are in tune with the structure of existing links, our method finds missing links that would immediately enhance Wikipedia's navigability. We leverage data sets of navigation paths collected through a Wikipedia-based human-computation game in which users must find a short path from a start to a target article by only clicking links encountered along the way. We harness human navigational traces to identify a set of candidates for missing links and then rank these candidates. Experiments show that our procedure identifies missing links of high quality.

Online crowdfunding platforms like DonorsChoose.org and Kick-starter allow specific projects to get funded by targeted contributions from a large number of people. Critical for the success of crowdfunding communities is recruitment and continued engagement of donors. With donor attrition rates above 70%, a significant challenge for online crowdfunding platforms as well as traditional offline non-profit organizations is the problem of donor retention. We present a large-scale study of millions of donors and donations on DonorsChoose.org, a crowdfunding platform for education projects. Studying an online crowdfunding platform allows for an unprecedented detailed view of how people direct their donations. We explore various factors impacting donor retention which allows us to identify different groups of donors and quantify their propensity to return for subsequent donations. We find that donors are more likely to return if they had a positive interaction with the receiver of the donation. We also show that this includes appropriate and timely recognition of their support as well as detailed communication of their impact. Finally, we discuss how our findings could inform steps to improve donor retention in crowdfunding communities and non-profit organizations.

Many of the world's most popular websites catalyze their growth through invitations from existing members. New members can then in turn issue invitations, and so on, creating cascades of member signups that can spread on a global scale. Although these diffusive invitation processes are critical to the popularity and growth of many websites, they have rarely been studied, and their properties remain elusive. For instance, it is not known how viral these cascades structures are, how cascades grow over time, or how diffusive growth affects the resulting distribution of member characteristics present on the site. In this paper, we study the diffusion of LinkedIn, an online professional network comprising over 332 million members, a large fraction of whom joined the site as part of a signup cascade. First we analyze the structural patterns of these signup cascades, and find them to be qualitatively different from previously studied information diffusion cascades. We also examine how signup cascades grow over time, and observe that diffusion via invitations on LinkedIn occurs over much longer timescales than are typically associated with other types of online diffusion. Finally, we connect the cascade structures with rich individual-level attribute data to investigate the interplay between the two. Using novel techniques to study the role of homophily in diffusion, we find striking differences between the local, edge-wise homophily and the global, cascade-level homophily we observe in our data, suggesting that signup cascades form surprisingly coherent groups of members.

Given the extremely large pool of events and stories available, media outlets need to focus on a subset of issues and aspects to convey to their audience. Outlets are often accused of exhibiting a systematic bias in this selection process, with different outlets portraying different versions of reality. However, in the absence of objective measures and empirical evidence, the direction and extent of systematicity remains widely disputed. In this paper we propose a framework based on quoting patterns for quantifying and characterizing the degree to which media outlets exhibit systematic bias. We apply this framework to a massive dataset of news articles spanning the six years of Obama's presidency and all of his speeches, and reveal that a systematic pattern does indeed emerge from the outlet's quoting behavior. Moreover, we show that this pattern can be successfully exploited in an unsupervised prediction setting, to determine which new quotes an outlet will select to broadcast. By encoding bias patterns in a low-rank space we provide an analysis of the structure of political media coverage. This reveals a latent media bias space that aligns surprisingly well with political ideology and outlet type. A linguistic analysis exposes striking differences across these latent dimensions, showing how the different types of media outlets portray different realities even when reporting on the same events. For example, outlets mapped to the mainstream conservative side of the latent space focus on quotes that portray a presidential persona disproportionately characterized by negativity.

An important part of finding information online involves clicking from page to page until an information need is fully satisfied. This is a complex task that can easily be frustrating and force users to give up prematurely. An empirical analysis of what makes users abandon click-based navigation tasks is hard, since most passively collected browsing logs do not specify the exact target page that a user was trying to reach. We propose to overcome this problem by using data collected via Wikispeedia, a Wikipedia-based human-computation game, in which users are asked to navigate from a start page to an explicitly given target page (both Wikipedia articles) by only tracing hyperlinks between Wikipedia articles. Our contributions are two-fold. First, by analyzing the differences between successful and abandoned navigation paths, we aim to understand what types of behavior are indicative of users giving up their navigation task. We also investigate how users make use of back clicks during their navigation. We find that users prefer backtracking to high-degree nodes that serve as landmarks and hubs for exploring the network of pages. Second, based on our analysis, we build statistical models for predicting whether a user will finish or abandon a navigation task, and if the next action will be a back click. Being able to predict these events is important as it can potentially help us design more human-friendly browsing interfaces and retain users who would otherwise have given up navigating a website.

Networks are a general language for representing relational information among objects. An effective way to model, reason about, and summarize networks, is to discover sets of nodes with common connectivity patterns. Such sets are commonly referred to as network communities. Research on network community detection has predominantly focused on identifying communities of densely connected nodes in undirected networks.

The Web has enabled one of the most visible recent developments in education---the deployment of massive open online courses. With their global reach and often staggering enrollments, MOOCs have the potential to become a major new mechanism for learning. Despite this early promise, however, MOOCs are still relatively unexplored and poorly understood.

On many social networking web sites such as Facebook and Twitter, resharing or reposting functionality allows users to share others' content with their own friends or followers. As content is reshared from user to user, large cascades of reshares can form. While a growing body of research has focused on analyzing and characterizing such cascades, a recent, parallel line of work has argued that the future trajectory of a cascade may be inherently unpredictable. In this work, we develop a framework for addressing cascade prediction problems. On a large sample of photo reshare cascades on Facebook, we find strong performance in predicting whether a cascade will continue to grow in the future. We find that the relative growth of a cascade becomes more predictable as we observe more of its reshares, that temporal and structural features are key predictors of cascade size, and that initially, breadth, rather than depth in a cascade is a better indicator of larger cascades. This prediction performance is robust in the sense that multiple distinct classes of features all achieve similar performance. We also discover that temporal features are predictive of a cascade's eventual shape. Observing independent cascades of the same content, we find that while these cascades differ greatly in size, we are still able to predict which ends up the largest.

In online social media systems users are not only posting, consuming, and resharing content, but also creating new and destroying existing connections in the underlying social network. While each of these two types of dynamics has individually been studied in the past, much less is known about the connection between the two. How does user information posting and seeking behavior interact with the evolution of the underlying social network structure? Here, we study ways in which network structure reacts to users posting and sharing content. We examine the complete dynamics of the Twitter information network, where users post and reshare information while they also create and destroy connections. We find that the dynamics of network structure can be characterized by steady rates of change, interrupted by sudden bursts. Information diffusion in the form of cascades of post re-sharing often creates such sudden bursts of new connections, which significantly change users' local network structure. We also explore the effect of the information content on the dynamics of the network and find evidence that the appearance of new topics and real-world events can lead to significant changes in edge creations and deletions. Lastly, we develop a model that quantifies the dynamics of the network and the occurrence of these bursts as a function of the information spreading through the network. The model can successfully predict which information diffusion events will lead to bursts in network dynamics.

Event sequences, such as patients' medical histories or users' sequences of product reviews, trace how individuals progress over time. Identifying common patterns, or progression stages, in such event sequences is a challenging task because not every individual follows the same evolutionary pattern, stages may have very different lengths, and individuals may progress at different rates. In this paper, we develop a model-based method for discovering common progression stages in general event sequences. We develop a generative model in which each sequence belongs to a class, and sequences from a given class pass through a common set of stages, where each sequence evolves at its own rate. We then develop a scalable algorithm to infer classes of sequences, while also segmenting each sequence into a set of stages. We evaluate our method on event sequences, ranging from patients' medical histories to online news and navigational traces from the Web. The evaluation shows that our methodology can predict future events in a sequence, while also accurately inferring meaningful progression stages, and effectively grouping sequences based on common progression patterns. More generally, our methodology allows us to reason about how event sequences progress over time, by discovering patterns and categories of temporal evolution in large-scale datasets of events.

One of the main organizing principles in real-world networks is that of network communities , where sets of nodes organize into densely linked clusters. Even though detection of such communities is of great interest, understanding the structure communities in large networks remains relatively limited. In particular, due to the unavailability of labeled ground-truth data, it was traditionally very hard to develop accurate models of network community structure. Here we use six large social, collaboration, and information networks where nodes explicitly state their ground-truth community memberships. For example, nodes in social networks join into explicitly defined interest based groups, and we use such groups as explicitly labeled ground-truth communities. We use such ground-truth communities to study their structural signatures by analyzing how ground-truth communities emerge in networks and how they overlap. We observe some surprising phenomena. First, ground-truth communities contain high-degree hub nodes that reside in community overlaps and link to most of the members of the community. Second, the overlaps of communities are more densely connected than the non-overlapping parts of communities. We show that this in contrast to the conventional wisdom that community overlaps are more sparsely connected than the non-overlapping parts themselves. We then show that many existing models of network communities do not capture dense community overlaps. This in turn means that most present models and community detection methods confuse overlaps as separate communities. In contrast, we present the community-affiliation graph model (AGM), a conceptual model of network community structure. We demonstrate that AGM reliably captures the overall structure of networks as well as the overlapping and hierarchical nature of network communities.

People's personal social networks are big and cluttered, and currently there is no good way to automatically organize them. Social networking sites allow users to manually categorize their friends into social circles (e.g., “circles” on Google+, and “lists” on Facebook and Twitter). However, circles are laborious to construct and must be manually updated whenever a user's network grows. In this article, we study the novel task of automatically identifying users' social circles. We pose this task as a multimembership node clustering problem on a user's ego network, a network of connections between her friends. We develop a model for detecting circles that combines network structure as well as user profile information. For each circle, we learn its members and the circle-specific user profile similarity metric. Modeling node membership to multiple circles allows us to detect overlapping as well as hierarchically nested circles. Experiments show that our model accurately identifies circles on a diverse set of data from Facebook, Google+, and Twitter, for all of which we obtain hand-labeled ground truth.

With massive amounts of data being generated and stored ubiquitously in every discipline and every aspect of our daily life, how to handle such big data poses many challenging issues to researchers in data and information systems. The participants of CIKM 2013 are active researchers on large scale data, information and knowledge management, from multiple disciplines, including database systems, data mining, information retrieval, human-computer interaction, and knowledge or information management.

Community detection algorithms are fundamental tools that allow us to uncover organizational principles in networks. When detecting communities, there are two possible sources of information one can use: the network structure, and the features and attributes of nodes. Even though communities form around nodes that have common edges and common attributes, typically, algorithms have only focused on one of these two data modalities: community detection algorithms traditionally focus only on the network structure, while clustering algorithms mostly consider only node attributes. In this paper, we develop Communities from Edge Structure and Node Attributes (CESNA), an accurate and scalable algorithm for detecting overlapping communities in networks with node attributes. CESNA statistically models the interaction between the network structure and the node attributes, which leads to more accurate community detection as well as improved robustness in the presence of noise in the network structure. CESNA has a linear runtime in the network size and is able to process networks an order of magnitude larger than comparable approaches. Last, CESNA also helps with the interpretation of detected communities by finding relevant node attributes for each community.

In an era of information overload, many people struggle to make sense of complex stories, such as presidential elections or economic reforms. We propose a methodology for creating structured summaries of information, which we call zoomable metro maps. Just as cartographic maps have been relied upon for centuries to help us understand our surroundings, metro maps can help us understand the information landscape.

In order to recommend products to users we must ultimately predict how a user will respond to a new product. To do so we must uncover the implicit tastes of each user as well as the properties of each product. For example, in order to predict whether a user will enjoy Harry Potter, it helps to identify that the book is about wizards, as well as the user's level of interest in wizardry. User feedback is required to discover these latent product and user dimensions. Such feedback often comes in the form of a numeric rating accompanied by review text. However, traditional methods often discard review text, which makes user and product latent dimensions difficult to interpret, since they ignore the very text that justifies a user's rating. In this paper, we aim to combine latent rating dimensions (such as those of latent-factor recommender systems) with latent review topics (such as those learned by topic models like LDA). Our approach has several advantages. Firstly, we obtain highly interpretable textual labels for latent rating dimensions, which helps us to `justify' ratings with text. Secondly, our approach more accurately predicts product ratings by harnessing the information present in review text; this is especially true for new products and users, who may have too few ratings to model their latent factors, yet may still provide substantial information from the text of even a single review. Thirdly, our discovered topics can be used to facilitate other tasks such as automated genre discovery, and to identify useful and representative reviews.

Diffusion of information, spread of rumors and infectious diseases are all instances of stochastic processes that occur over the edges of an underlying network. Many times networks over which contagions spread are unobserved, and such networks are often dynamic and change over time. In this paper, we investigate the problem of inferring dynamic networks based on information diffusion data. We assume there is an unobserved dynamic network that changes over time, while we observe the results of a dynamic process spreading over the edges of the network. The task then is to infer the edges and the dynamics of the underlying network.

Network communities represent basic structures for understanding the organization of real-world networks. A community (also referred to as a module or a cluster) is typically thought of as a group of nodes with more connections amongst its members than between its members and the remainder of the network. Communities in networks also overlap as nodes belong to multiple clusters at once. Due to the difficulties in evaluating the detected communities and the lack of scalable algorithms, the task of overlapping community detection in large networks largely remains an open problem.

An increasingly common feature of online communities and social media sites is a mechanism for rewarding user achievements based on a system of badges. Badges are given to users for particular contributions to a site, such as performing a certain number of actions of a given type. They have been employed in many domains, including news sites like the Huffington Post, educational sites like Khan Academy, and knowledge-creation sites like Wikipedia and Stack Overflow. At the most basic level, badges serve as a summary of a user's key accomplishments; however, experience with these sites also shows that users will put in non-trivial amounts of work to achieve particular badges, and as such, badges can act as powerful incentives. Thus far, however, the incentive structures created by badges have not been well understood, making it difficult to deploy badges with an eye toward the incentives they are likely to create.

Vibrant online communities are in constant flux. As members join and depart, the interactional norms evolve, stimulating further changes to the membership and its social dynamics. Linguistic change --- in the sense of innovation that becomes accepted as the norm --- is essential to this dynamic process: it both facilitates individual expression and fosters the emergence of a collective identity.

Recommending products to consumers means not only understanding their tastes, but also understanding their level of experience. For example, it would be a mistake to recommend the iconic film Seven Samurai simply because a user enjoys other action movies; rather, we might conclude that they will eventually enjoy it---once they are ready. The same is true for beers, wines, gourmet foods---or any products where users have acquired tastes: the `best' products may not be the most 'accessible'. Thus our goal in this paper is to recommend products that a user will enjoy now, while acknowledging that their tastes may have changed over time, and may change again in the future. We model how tastes change due to the very act of consuming more products---in other words, as users become more experienced. We develop a latent factor recommendation system that explicitly accounts for each user's level of experience. We find that such a model not only leads to better recommendations, but also allows us to study the role of user experience and expertise on a novel dataset of fifteen million beer, wine, food, and movie reviews.

The real-time information on news sites, blogs and social networking sites changes dynamically and spreads rapidly through the Web. Developing methods for handling such information at a massive scale requires that we think about how information content varies over time, how it is transmitted, and how it mutates as it spreads.

Most online reviews consist of plain-text feedback together with a single numeric score. However, understanding the multiple `aspects' that contribute to users' ratings may help us to better understand their individual preferences. For example, a user's impression of an audio book presumably depends on aspects such as the story and the narrator, and knowing their opinions on these aspects may help us to recommend better products. In this paper, we build models for rating systems in which such dimensions are explicit, in the sense that users leave separate ratings for each aspect of a product. By introducing new corpora consisting of five million reviews, rated with between three and six aspects, we evaluate our models on three prediction tasks: First, we uncover which parts of a review discuss which of the rated aspects. Second, we summarize reviews by finding the sentences that best explain a user's rating. Finally, since aspect ratings are optional in many of the datasets we consider, we recover ratings that are missing from a user's evaluation. Our model matches state-of-the-art approaches on existing small-scale datasets, while scaling to the real-world datasets we introduce. Moreover, our model is able to `disentangle' content and sentiment words: we automatically learn content words that are indicative of a particular aspect as well as the aspect-specific sentiment words that are indicative of a particular rating.

In networks, contagions such as information, purchasing behaviors, and diseases, spread and diffuse from node to node over the edges of the network. Moreover, in real-world scenarios multiple contagions spread through the network simultaneously. These contagions not only propagate at the same time but they also interact and compete with each other as they spread over the network. While traditional empirical studies and models of diffusion consider individual contagions as independent and thus spreading in isolation, we study how different contagions interact with each other as they spread through the network. We develop a statistical model that allows for competition as well as cooperation of different contagions in information diffusion. Competing contagions decrease each other's probability of spreading, while cooperating contagions help each other in being adopted throughout the network. We evaluate our model on 18,000 contagions simultaneously spreading through the Twitter network. Our model learns how different contagions interact with each other and then uses these interactions to more accurately predict the diffusion of a contagion through the network. Moreover, the model also provides a compelling hypothesis for the principles that govern content interaction in information diffusion. Most importantly, we find very strong effects of interactions between contagions. Interactions cause a relative change in the spreading probability of a contagion by 71% on the average.

Nodes in real-world networks organize into densely linked communities where edges appear with high concentration among the members of the community. Identifying such communities of nodes has proven to be a challenging task mainly due to a plethora of definitions of a community, intractability of algorithms, issues with evaluation and the lack of a reliable gold-standard ground-truth. In this paper we study a set of 230 large real-world social, collaboration and information networks where nodes explicitly state their group memberships. For example, in social networks nodes explicitly join various interest based social groups. We use such groups to define a reliable and robust notion of ground-truth communities. We then propose a methodology which allows us to compare and quantitatively evaluate how different structural definitions of network communities correspond to ground-truth communities. We choose 13 commonly used structural definitions of network communities and examine their sensitivity, robustness and performance in identifying the ground-truth. We show that the 13 structural definitions are heavily correlated and naturally group into four classes. We find that two of these definitions, Conductance and Triad-participation-ratio, consistently give the best performance in identifying ground-truth communities. We also investigate a task of detecting communities given a single seed node. We extend the local spectral clustering algorithm into a heuristic parameter-free community detection method that easily scales to networks with more than hundred million nodes. The proposed method achieves 30% relative improvement over current local clustering methods.

One of the main organizing principles in real-world networks is that of network communities, where sets of nodes organize into densely linked clusters. Communities in networks often overlap as nodes can belong to multiple communities at once. Identifying such overlapping communities is crucial for the understanding the structure as well as the function of real-world networks. Even though community structure in networks has been widely studied in the past, practically all research makes an implicit assumption that overlaps between communities are less densely connected than the non-overlapping parts themselves. Here we validate this assumption on 6 large scale social, collaboration and information networks where nodes explicitly state their community memberships. By examining such ground-truth communities we find that the community overlaps are more densely connected than the non-overlapping parts, which is in sharp contrast to the conventional wisdom that community overlaps are more sparsely connected than the communities themselves. Practically all existing community detection methods fail to detect communities with dense overlaps. We propose Community-Affiliation Graph Model, a model-based community detection method that builds on bipartite node-community affiliation networks. Our method successfully captures overlapping, non-overlapping as well as hierarchically nested communities, and identifies relevant communities more accurately than the state-of-the-art methods in networks ranging from biological to social and information networks.

Question answering (Q&A) websites are now large repositories of valuable knowledge. While most Q&A sites were initially aimed at providing useful answers to the question asker, there has been a marked shift towards question answering as a community-driven knowledge creation process whose end product can be of enduring value to a broad audience. As part of this shift, specific expertise and deep knowledge of the subject at hand have become increasingly important, and many Q&A sites employ voting and reputation mechanisms as centerpieces of their design to help users identify the trustworthiness and accuracy of the content.

Social networks play a fundamental role in the diffusion of information. However, there are two different ways of how information reaches a person in a network. Information reaches us through connections in our social networks, as well as through the influence external out-of-network sources, like the mainstream media. While most present models of information adoption in networks assume information only passes from a node to node via the edges of the underlying network, the recent availability of massive online social media data allows us to study this process in more detail.

There are many settings in which users of a social media application provide evaluations of one another. In a variety of domains, mechanisms for evaluation allow one user to say whether he or she trusts another user, or likes the content they produced, or wants to confer special levels of authority or responsibility on them. Earlier work has studied how the relative status between two users - that is, their comparative levels of status in the group - affects the types of evaluations that one user gives to another.

We pose a fundamental question in understanding how to identify and design successful communities: What factors predict whether a community will grow and survive in the long term? Social scientists have addressed this question extensively by analyzing offline groups which endeavor to attract new members, such as social movements, finding that new individuals are influenced strongly by their ties to members of the group. As a result, prior work on the growth of communities has treated growth primarily as a diffusion processes, leading to findings about group evolution which can be difficult to explain. The proliferation of online social networks and communities, however, has created new opportunities to study, at a large scale and with very fine resolution, the mechanisms which lead to the formation, growth, and demise of online groups.

Navigating information spaces is an essential part of our everyday lives, and in order to design efficient and user-friendly information systems, it is important to understand how humans navigate and find the information they are looking for. We perform a large-scale study of human wayfinding, in which, given a network of links between the concepts of Wikipedia, people play a game of finding a short path from a given start to a given target concept by following hyperlinks. What distinguishes our setup from other studies of human Web-browsing behavior is that in our case people navigate a graph of connections between concepts, and that the exact goal of the navigation is known ahead of time. We study more than 30,000 goal-directed human search paths and identify strategies people use when navigating information spaces. We find that human wayfinding, while mostly very efficient, differs from shortest paths in characteristic ways. Most subjects navigate through high-degree hubs in the early phase, while their search is guided by content features thereafter. We also observe a trade-off between simplicity and efficiency: conceptually simple solutions are more common but tend to be less efficient than more complex ones. Finally, we consider the task of predicting the target a user is trying to reach. We design a model and an efficient learning algorithm. Such predictive models of human wayfinding can be applied in intelligent browsing interfaces.

Information diffusion and virus propagation are fundamental processes taking place in networks. While it is often possible to directly observe when nodes become infected with a virus or publish the information, observing individual transmissions (who infects whom, or who influences whom) is typically very difficult. Furthermore, in many applications, the underlying network over which the diffusions and propagations spread is actually unobserved. We tackle these challenges by developing a method for tracing paths of diffusion and influence through networks and inferring the networks over which contagions propagate. Given the times when nodes adopt pieces of information or become infected, we identify the optimal network that best explains the observed infection times. Since the optimization problem is NP-hard to solve exactly, we develop an efficient approximation algorithm that scales to large datasets and finds provably near-optimal networks. We demonstrate the effectiveness of our approach by tracing information diffusion in a set of 170 million blogs and news articles over a one year period to infer how information flows through the online media space. We find that the diffusion network of news for the top 1,000 media sites and blogs tends to have a core-periphery structure with a small set of core media sites that diffuse information to the rest of the Web. These sites tend to have stable circles of influence with more general news media sites acting as connectors between them.

Even though human movement and mobility patterns have a high degree of freedom and variation, they also exhibit structural patterns due to geographic and social constraints. Using cell phone location data, as well as data from two online location-based social networks, we aim to understand what basic laws govern human motion and dynamics. We find that humans experience a combination of periodic movement that is geographically limited and seemingly random jumps correlated with their social networks. Short-ranged travel is periodic both spatially and temporally and not effected by the social network structure, while long-distance travel is more influenced by social network ties. We show that social relationships can explain about 10% to 30% of all human movement, while periodic behavior explains 50% to 70%. Based on our findings, we develop a model of human mobility that combines periodic short range movements with travel due to the social network structure. We show that our model reliably predicts the locations and dynamics of future human movement and gives an order of magnitude better performance than present models of human mobility.

While the social and information networks have become ubiquitous, the challenge of collecting complete network data still persists. Many times the collected network data is incomplete with nodes and edges missing. Commonly, only a part of the network can be observed and we would like to infer the unobserved part of the network. We address this issue by studying the Network Completion Problem: Given a network with missing nodes and edges, can we complete the missing part? We cast the problem in the Expectation Maximization (EM) framework where we use the observed part of the network to fit a model of network structure, and then we estimate the missing part of the network using the model, re-estimate the parameters and so on. We combine the EM algorithm with the Kronecker graphs model and design a scalable Metropolized Gibbs sampling approach that allows for the estimation of the model parameters as well as the inference about missing nodes and edges of the network. Experiments on synthetic and several real-world networks show that our approach can effectively recover the network even when about half of the nodes in the network are missing. Our algorithm outperforms not only classical link-prediction approaches but also the state of the art Stochastic block modeling approach. Furthermore, our algorithm easily scales to networks with tens of thousands of nodes.

Predicting the occurrence of links is a fundamental problem in networks. In the link prediction problem we are given a snapshot of a network and would like to infer which interactions among existing members are likely to occur in the near future or which existing interactions are we missing. Although this problem has been extensively studied, the challenge of how to effectively combine the information from the network structure with rich node and edge attribute data remains largely open.

Transmission of infectious diseases, propagation of information, and spread of ideas and influence through social networks are all examples of diffusion. In such cases we say that a contagion spreads through the network, a process that can be modeled by a cascade graph. Studying cascades and network diffusion is challenging due to missing data. Even a single missing observation in a sequence of propagation events can significantly alter our inferences about the diffusion process.

Online content exhibits rich temporal dynamics, and diverse realtime user generated content further intensifies this process. However, temporal patterns by which online content grows and fades over time, and by which different pieces of content compete for attention remain largely unexplored. We study temporal patterns associated with online content and how the content’s popularity grows and fades over time. The attention that content receives on the Web varies depending on many factors and occurs on very different time scales and at different resolutions. In order to uncover the temporal dynamics of online content we formulate a time series clustering problem using a similarity metric that is invariant to scaling and shifting. We develop the K-Spectral Centroid (K-SC) clustering algorithm that effectively finds cluster centroids with our similarity measure. By applying an adaptive wavelet-based incremental approach to clustering, we scale K-SC to large data sets. We demonstrate our approach on two massive datasets: a set of 580 million Tweets, and a set of 170 million blog posts and news media articles. We find that K-SC outperforms the K-means clustering algorithm in finding distinct shapes of time series. Our analysis shows that there are six main temporal shapes of attention of online content. We also present a simple model that reliably predicts the shape of attention by using information about only a small number of participants. Our analyses offer insight into common temporal patterns of the content on the Web and broaden the understanding of the dynamics of human attention.

Online peer-to-peer (P2P) lending services are a new type of social platform that enables individuals borrow and lend money directly from one to another. In this paper, we study the dynamics of bidding behavior in a P2P loan auction website, Prosper.com. We investigate the change of various attributes of loan requesting listings over time, such as the interest rate and the number of bids. We observe that there is herding behavior during bidding, and for most of the listings, the numbers of bids they receive reach spikes at very similar time points. We explain these phenomena by showing that there are economic and social factors that lenders take into account when deciding to bid on a listing. We also observe that the profits the lenders make are tied with their bidding preferences. Finally, we build a model based on the temporal progression of the bidding, that reliably predicts the success of a loan request listing, as well as whether a loan will be paid back or not.

Given large, multimillion-node graphs (e.g., Facebook, Web-crawls, etc.), how do they evolve over time? How are they connected? What are the central nodes and the outliers? In this article we define the Radius plot of a graph and show how it can answer these questions. However, computing the Radius plot is prohibitively expensive for graphs reaching the planetary scale. There are two major contributions in this article: (a) We propose HADI (HAdoop DIameter and radii estimator), a carefully designed and fine-tuned algorithm to compute the radii and the diameter of massive graphs, that runs on the top of the Hadoop / MapReduce system, with excellent scale-up on the number of available machines (b) We run HADI on several real world datasets including YahooWeb (6B edges, 1/8 of a Terabyte), one of the largest public graphs ever analyzed. Thanks to HADI, we report fascinating patterns on large networks, like the surprisingly small effective diameter, the multimodal/bimodal shape of the Radius plot, and its palindrome motion over time.

Social media forms a central domain for the production and dissemination of real-time information. Even though such flows of information have traditionally been thought of as diffusion processes over social networks, the underlying phenomena are the result of a complex web of interactions among numerous participants. Here we develop the Linear Influence Model where rather than requiring the knowledge of the social network and then modeling the diffusion by predicting which node will influence which other nodes in the network, we focus on modeling the global influence of a node on the rate of diffusion through the (implicit) network. We model the number of newly infected nodes as a function of which other nodes got infected in the past. For each node we estimate an influence function that quantifies how many subsequent infections can be attributed to the influence of that node over time. A nonparametric formulation of the model leads to a simple least squares problem that can be solved on large datasets. We validate our model on a set of 500 million tweets and a set of 170 million news articles and blog posts. We show that the Linear Influence Model accurately models influences of nodes and reliably predicts the temporal dynamics of information diffusion. We find that patterns of influence of individual participants differ significantly depending on the type of the node and the topic of the information.

Information diffusion and virus propagation are fundamental processes talking place in networks. While it is often possible to directly observe when nodes become infected, observing individual transmissions (i.e., who infects whom or who influences whom) is typically very difficult. Furthermore, in many applications, the underlying network over which the diffusions and propagations spread is actually unobserved. We tackle these challenges by developing a method for tracing paths of diffusion and influence through networks and inferring the networks over which contagions propagate. Given the times when nodes adopt pieces of information or become infected, we identify the optimal network that best explains the observed infection times. Since the optimization problem is NP-hard to solve exactly, we develop an efficient approximation algorithm that scales to large datasets and in practice gives provably near-optimal performance. We demonstrate the effectiveness of our approach by tracing information cascades in a set of 170 million blogs and news articles over a one year period to infer how information flows through the online media space. We find that the diffusion network of news tends to have a core-periphery structure with a small set of core media sites that diffuse information to the rest of the Web. These sites tend to have stable circles of influence with more general news media sites acting as connectors between them.

Given large, multi-million node graphs (e.g., FaceBook, web-crawls, etc.), how do they evolve over time? How are they connected? What are the central nodes and the outliers of the graphs? We show that the Radius Plot (pdf of node radii) can answer these questions. However, computing the Radius Plot is prohibitively expensive for graphs reaching the planetary scale. There are two major contributions in this paper: (a) We propose HADI (HAdoop DIameter and radii estimator), a carefully designed and fine-tuned algorithm to compute the diameter of massive graphs, that runs on the top of the Hadoop/MapReduce system, with excellent scale-up on the number of available machines (b) We run HADI on several real world datasets including YahooWeb (6B edges, 1/8 of a Terabyte), one of the largest public graphs ever analyzed. Thanks to HADI, we report fascinating patterns on large networks, like the surprisingly small effective diameter, the multi-modal/bi-modal shape of the Radius Plot, and its palindrome motion over time.

We study online social networks in which relationships can be either positive (indicating relations such as friendship) or negative (indicating relations such as opposition or antagonism). Such a mix of positive and negative links arise in a variety of online settings; we study datasets from Epinions, Slashdot and Wikipedia. We find that the signs of links in the underlying social networks can be predicted with high accuracy, using models that generalize across this diverse range of sites. These models provide insight into some of the fundamental principles that drive the formation of signed links in networks, shedding light on theories of balance and status from social psychology; they also suggest social computing applications by which the attitude of one user toward another can be estimated from evidence provided by their relationships with other members of the surrounding social network.

Detecting clusters or communities in large real-world graphs such as large social or information networks is a problem of considerable interest. In practice, one typically chooses an objective function that captures the intuition of a network cluster as set of nodes with better internal connectivity than external connectivity, and then one applies approximation algorithms or heuristics to extract sets of nodes that are related to the objective function and that look like good communities for the application of interest.In this paper, we explore a range of network community detection methods in order to compare them and to understand their relative performance and the systematic biases in the clusters they identify. We evaluate several common objective functions that are used to formalize the notion of a network community, and we examine several different classes of approximation algorithms that aim to optimize such objective functions. In addition, rather than simply fixing an objective and asking for an approximation to the best cluster of any size, we consider a size-resolved version of the optimization problem. Considering community quality as a function of its size provides a much finer lens with which to examine community detection algorithms, since objective functions and approximation algorithms often have non-obvious size-dependent behavior.

Tracking new topics, ideas, and "memes" across the Web has been an issue of considerable interest. Recent work has developed methods for tracking topic shifts over long time scales, as well as abrupt spikes in the appearance of particular named entities. However, these approaches are less well suited to the identification of content that spreads widely and then fades over time scales on the order of days - the time scale at which we perceive news and events.

Graphs appear in several settings, like social networks, recommendation systems, computer communication networks, gene/protein biological networks, among others. A deep, recurring question is “What do real graphs look like?” That is, how can we separate real ones from synthetic or real graphs with masked portions? The main contribution of this paper is ShatterPlots, a simple and powerful algorithm to extract patterns from real graphs that help us spot fake/masked graphs. The idea is to shatter a graph, by deleting edges, force it to reach a critical (“Shattering”) point, and study the properties at that point. One of the most striking patterns is the “30-per-cent“: at the Shattering point, all real and synthetic graphs have about 30% more nodes than edges. One of our most discriminative patterns is the “NodeShatteringRatio“, which can almost perfectly separate the real graphs from the synthetic ones of our extensive collection. Additional contributions of this paper are (a) the careful, scalable design of the algorithm, which requires only O(E) time, (b) extensive experiments in a large collection of graphs (19 in total), with up to hundreds of thousands of nodes and million edges, and (c) a wealth of observations and patterns, which show how to distinguish synthetic or masked graphs from real ones.

Given publication titles and authors, what can we say about the evolution of scientific topics and communities over time? Which communities shrunk, which emerged, and which split, over time? And, when in time were the turning points? We propose TimeFall, which can automatically answer these questions given a social network/graph that evolves over time. The main novelty of the proposed approach is that it needs no user-defined parameters, relying instead on the principle of minimum description length (MDL), to extract the communities, and to find good cut-points in time when communities change abruptly: a cut-point is good, if it leads to shorter data description. We illustrate our algorithm on synthetic and large real datasets, and we show that the results of the TimeFall agree with human intuition.

By now, online social networks have become an indispensable part of both online and offline lives of human beings. A large fraction of time spent online by a user is directly influence by the social networks to which he/she belongs. This calls for a deeper examination of social networks as large-scale dynamic objects that foster efficient person-person interaction.

We present a detailed study of network evolution by analyzing four large online social networks with full temporal information about node and edge arrivals. For the first time at such a large scale, we study individual node arrival and edge creation processes that collectively lead to macroscopic properties of networks. Using a methodology based on the maximum-likelihood principle, we investigate a wide variety of network formation strategies, and show that edge locality plays a critical role in evolution of networks. Our findings supplement earlier network models based on the inherently non-local preferential attachment.

We analyze a massive social network, gathered from the records of a large mobile phone operator, with more than a million users and tens of millions of calls. We examine the distributions of the number of phone calls per customer; the total talk minutes per customer; and the distinct number of calling partners per customer. We find that these distributions are skewed, and that they significantly deviate from what would be expected by power-law and lognormal distributions.

We present a study of anonymized data capturing a month of high-level communication activities within the whole of the Microsoft Messenger instant-messaging system. We examine characteristics and patterns that emerge from the collective dynamics of large numbers of people, rather than the actions and characteristics of individuals. The dataset contains summary properties of 30 billion conversations among 240 million people. From the data, we construct a communication graph with 180 million nodes and 1.3 billion undirected edges, creating the largest social network constructed and analyzed to date. We report on multiple aspects of the dataset and synthesized graph. We find that the graph is well-connected and robust to node removal. We investigate on a planetary-scale the oft-cited report that people are separated by "six degrees of separation" and find that the average path length among Messenger users is 6.6. We find that people tend to communicate more with each other when they have similar age, language, and location, and that cross-gender conversations are both more frequent and of longer duration than conversations with the same gender.

A large body of work has been devoted to identifying community structure in networks. A community is often though of as a set of nodes that has more connections between its members than to the remainder of the network. In this paper, we characterize as a function of size the statistical and structural properties of such sets of nodes. We define the network community profile plot, which characterizes the "best" possible community - according to the conductance measure - over a wide range of size scales, and we study over 70 large sparse real-world networks taken from a wide range of application domains. Our results suggest a significantly more refined picture of community structure in large real-world networks than has been appreciated previously.

Given a water distribution network, where should we place sensors toquickly detect contaminants? Or, which blogs should we read to avoid missing important stories?.

How do blogs cite and influence each other? How do such links evolve? Does the popularity of old blog posts drop exponentially with time? These are some of the questions that we address in this work. Blogs (weblogs) have become an important medium of information because of their timely publication, ease of use, and wide availability. In fact, they often make headlines, by discussing and discovering evidence about political events and facts. Often blogs link to one another, creating a publicly available record of how information and influence spreads through an underlying social network. Aggregating links from several blog posts creates a directed graph which we analyze to discover the patterns of information propagation in blogspace, and thereby understand the underlying social network. Here we report some surprising findings of the blog linking and information propagation structure, after we analyzed one of the largest available datasets, with 45,000 blogs and ≈ 2.2 million blog-postings. Our analysis also sheds light on how rumors, viruses, and ideas propagate over social and computer networks.

Graphical relationships among Web pages have been exploited inmethods for ranking search results. To date, specific graphicalproperties have been used in these analyses. We introduce a WebProjection methodology that generalizes prior efforts of graphicalrelationships of the web in several ways. With the approach, wecreate subgraphs by projecting sets of pages and domains onto thelarger web graph, and then use machine learning to constructpredictive models that consider graphical properties as evidence. Wedescribe the method and then present experiments that illustrate theconstruction of predictive models of search result quality and userquery reformulation.

How do real graphs evolve over time? What are normal growth patterns in social, technological, and information networks? Many studies have discovered patterns in static graphs , identifying properties in a single snapshot of a large network or in a very small number of snapshots; these include heavy tails for in- and out-degree distributions, communities, small-world phenomena, and others. However, given the lack of information about network evolution over long periods, it has been hard to convert these findings into statements about trends over time. Here we study a wide range of real graphs, and we observe some surprising phenomena. First, most of these graphs densify over time with the number of edges growing superlinearly in the number of nodes. Second, the average distance between nodes often shrinks over time in contrast to the conventional wisdom that such distance parameters should increase slowly as a function of the number of nodes (like O (log n ) or O (log(log n )). Existing graph generation models do not exhibit these types of behavior even at a qualitative level. We provide a new graph generator, based on a forest fire spreading process that has a simple, intuitive justification, requires very few parameters (like the flammability of nodes), and produces graphs exhibiting the full range of properties observed both in prior work and in the present study. We also notice that the forest fire model exhibits a sharp transition between sparse graphs and graphs that are densifying. Graphs with decreasing distance between the nodes are generated around this transition point. Last, we analyze the connection between the temporal evolution of the degree distribution and densification of a graph. We find that the two are fundamentally related. We also observe that real networks exhibit this type of relation between densification and the degree distribution.

We present an analysis of a person-to-person recommendation network, consisting of 4 million people who made 16 million recommendations on half a million products. We observe the propagation of recommendations and the cascade sizes, which we explain by a simple stochastic model. We analyze how user behavior varies within user communities defined by a recommendation network. Product purchases follow a ‘long tail’ where a significant share of purchases belongs to rarely sold items. We establish how the recommendation network grows over time and how effective it is from the viewpoint of the sender and receiver of the recommendations. While on average recommendations are not very effective at inducing purchases and do not spread very far, we present a model that successfully identifies communities, product, and pricing categories for which viral marketing seems to be very effective.

Given a huge real graph, how can we derive a representative sample? There are many known algorithms to compute interesting measures (shortest paths, centrality, betweenness, etc.), but several of them become impractical for large graphs. Thus graph sampling is essential.The natural questions to ask are (a) which sampling method to use, (b) how small can the sample size be, and (c) how to scale up the measurements of the sample (e.g., the diameter), to get estimates for the large graph. The deeper, underlying question is subtle: how do we measure success?.We answer the above questions, and test our answers by thorough experiments on several, diverse datasets, spanning thousands nodes and edges. We consider several sampling methods, propose novel methods to check the goodness of sampling, and develop a set of scaling laws that describe relations between the properties of the original and the sample.In addition to the theoretical contributions, the practical conclusions from our work are: Sampling strategies based on edge selection do not perform well; simple uniform random node selection performs surprisingly well. Overall, best performing methods are the ones based on random-walks and "forest fire"; they match very accurately both static as well as evolutionary graph patterns, with sample sizes down to about 15% of the original graph.

How do real graphs evolve over time? What are "normal" growth patterns in social, technological, and information networks? Many studies have discovered patterns in static graphs, identifying properties in a single snapshot of a large network, or in a very small number of snapshots; these include heavy tails for in- and out-degree distributions, communities, small-world phenomena, and others. However, given the lack of information about network evolution over long periods, it has been hard to convert these findings into statements about trends over time.Here we study a wide range of real graphs, and we observe some surprising phenomena. First, most of these graphs densify over time, with the number of edges growing super-linearly in the number of nodes. Second, the average distance between nodes often shrinks over time, in contrast to the conventional wisdom that such distance parameters should increase slowly as a function of the number of nodes (like O(log n) or O(log(log n)).Existing graph generation models do not exhibit these types of behavior, even at a qualitative level. We provide a new graph generator, based on a "forest fire" spreading process, that has a simple, intuitive justification, requires very few parameters (like the "flammability" of nodes), and produces graphs exhibiting the full range of properties observed both in prior work and in the present study.