arXiv subjects

Yuxiao Dong

Publications and source records attributed to Yuxiao Dong.

At least 127 records · Page 7Linked to original sources

arXiv2020

GCC: Graph Contrastive Coding for Graph Neural Network Pre-Training

Graph representation learning has emerged as a powerful technique for addressing real-world problems. Various downstream graph learning tasks have benefited from its recent developments, such as node classification, similarity search, and graph classification. However, prior arts on graph representation learning focus on domain specific problems and train a dedicated model for each graph dataset, which is usually non-transferable to out-of-domain data. Inspired by the recent advances in pre-training from natural language processing and computer vision, we design Graph Contrastive Coding (GCC) -- a self-supervised graph neural network pre-training framework -- to capture the universal network topological properties across multiple networks. We design GCC's pre-training task as subgraph instance discrimination in and across networks and leverage contrastive learning to empower graph neural networks to learn the intrinsic and transferable structural representations. We conduct extensive experiments on three graph learning tasks and ten graph datasets. The results show that GCC pre-trained on a collection of diverse datasets can achieve competitive or better performance to its task-specific and trained-from-scratch counterparts. This suggests that the pre-training and fine-tuning paradigm presents great potential for graph representation learning.

cs.LG↗

arXiv2020

GPT-GNN: Generative Pre-Training of Graph Neural Networks

Graph neural networks (GNNs) have been demonstrated to be powerful in modeling graph-structured data. However, training GNNs usually requires abundant task-specific labeled data, which is often arduously expensive to obtain. One effective way to reduce the labeling effort is to pre-train an expressive GNN model on unlabeled data with self-supervision and then transfer the learned model to downstream tasks with only a few labels. In this paper, we present the GPT-GNN framework to initialize GNNs by generative pre-training. GPT-GNN introduces a self-supervised attributed graph generation task to pre-train a GNN so that it can capture the structural and semantic properties of the graph. We factorize the likelihood of the graph generation into two components: 1) Attribute Generation and 2) Edge Generation. By modeling both components, GPT-GNN captures the inherent dependency between node attributes and graph structure during the generative process. Comprehensive experiments on the billion-scale Open Academic Graph and Amazon recommendation data demonstrate that GPT-GNN significantly outperforms state-of-the-art GNN models without pre-training by up to 9.1% across various downstream tasks.

cs.LG↗

arXiv2020

Heterogeneous Graph Transformer

Recent years have witnessed the emerging success of graph neural networks (GNNs) for modeling structured data. However, most GNNs are designed for homogeneous graphs, in which all nodes and edges belong to the same types, making them infeasible to represent heterogeneous structures. In this paper, we present the Heterogeneous Graph Transformer (HGT) architecture for modeling Web-scale heterogeneous graphs. To model heterogeneity, we design node- and edge-type dependent parameters to characterize the heterogeneous attention over each edge, empowering HGT to maintain dedicated representations for different types of nodes and edges. To handle dynamic heterogeneous graphs, we introduce the relative temporal encoding technique into HGT, which is able to capture the dynamic structural dependency with arbitrary durations. To handle Web-scale graph data, we design the heterogeneous mini-batch graph sampling algorithm---HGSampling---for efficient and scalable training. Extensive experiments on the Open Academic Graph of 179 million nodes and 2 billion edges show that the proposed HGT model consistently outperforms all the state-of-the-art GNN baselines by 9%--21% on various downstream tasks.

cs.LG↗

arXiv2019

NetSMF: Large-Scale Network Embedding as Sparse Matrix Factorization

We study the problem of large-scale network embedding, which aims to learn latent representations for network mining applications. Previous research shows that 1) popular network embedding benchmarks, such as DeepWalk, are in essence implicitly factorizing a matrix with a closed form, and 2)the explicit factorization of such matrix generates more powerful embeddings than existing methods. However, directly constructing and factorizing this matrix---which is dense---is prohibitively expensive in terms of both time and space, making it not scalable for large networks. In this work, we present the algorithm of large-scale network embedding as sparse matrix factorization (NetSMF). NetSMF leverages theories from spectral sparsification to efficiently sparsify the aforementioned dense matrix, enabling significantly improved efficiency in embedding learning. The sparsified matrix is spectrally close to the original dense one with a theoretically bounded approximation error, which helps maintain the representation power of the learned embeddings. We conduct experiments on networks of various scales and types. Results show that among both popular benchmarks and factorization based methods, NetSMF is the only method that achieves both high efficiency and effectiveness. We show that NetSMF requires only 24 hours to generate effective embeddings for a large-scale academic collaboration network with tens of millions of nodes, while it would cost DeepWalk months and is computationally infeasible for the dense matrix factorization solution. The source code of NetSMF is publicly available (https://github.com/xptree/NetSMF).

cs.SI↗

arXiv2018

DeepInf: Social Influence Prediction with Deep Learning

Social and information networking activities such as on Facebook, Twitter, WeChat, and Weibo have become an indispensable part of our everyday life, where we can easily access friends' behaviors and are in turn influenced by them. Consequently, an effective social influence prediction for each user is critical for a variety of applications such as online recommendation and advertising. Conventional social influence prediction approaches typically design various hand-crafted rules to extract user- and network-specific features. However, their effectiveness heavily relies on the knowledge of domain experts. As a result, it is usually difficult to generalize them into different domains. Inspired by the recent success of deep neural networks in a wide range of computing applications, we design an end-to-end framework, DeepInf, to learn users' latent feature representation for predicting social influence. In general, DeepInf takes a user's local network as the input to a graph neural network for learning her latent social representation. We design strategies to incorporate both network structures and user-specific features into convolutional neural and attention networks. Extensive experiments on Open Academic Graph, Twitter, Weibo, and Digg, representing different types of social and information networks, demonstrate that the proposed end-to-end model, DeepInf, significantly outperforms traditional feature engineering-based approaches, suggesting the effectiveness of representation learning for social applications.

cs.SI↗

arXiv2018

Collaboration Diversity and Scientific Impact

The shift from individual effort to collaborative output has benefited science, with scientific work pursued collaboratively having increasingly led to more highly impactful research than that pursued individually. However, understanding of how the diversity of a collaborative team influences the production of knowledge and innovation is sorely lacking. Here, we study this question by breaking down the process of scientific collaboration of 32.9 million papers over the last five decades. We find that the probability of producing a top-cited publication increases as a function of the diversity of a team of collaborators---namely, the distinct number of institutions represented by the team. We discover striking phenomena where a smaller, yet more diverse team is more likely to generate highly innovative work than a relatively larger team within one institution. We demonstrate that the synergy of collaboration diversity is universal across different generations, research fields, and tiers of institutions and individual authors. Our findings suggest that collaboration diversity strongly and positively correlates with the production of scientific innovation, giving rise to the potential revolution of the policies used by funding agencies and authorities to fund research projects, and broadly the principles used to organize teams, organizations, and societies.

cs.DL↗

arXiv2018

Neural Tensor Factorization

Neural collaborative filtering (NCF) and recurrent recommender systems (RRN) have been successful in modeling user-item relational data. However, they are also limited in their assumption of static or sequential modeling of relational data as they do not account for evolving users' preference over time as well as changes in the underlying factors that drive the change in user-item relationship over time. We address these limitations by proposing a Neural Tensor Factorization (NTF) model for predictive tasks on dynamic relational data. The NTF model generalizes conventional tensor factorization from two perspectives: First, it leverages the long short-term memory architecture to characterize the multi-dimensional temporal interactions on relational data. Second, it incorporates the multi-layer perceptron structure for learning the non-linearities between different latent factors. Our extensive experiments demonstrate the significant improvement in rating prediction and link prediction on dynamic relational data by our NTF model over both neural network based factorization models and other traditional methods.

cs.LG↗

arXiv2018

Network Embedding as Matrix Factorization: Unifying DeepWalk, LINE, PTE, and node2vec

Since the invention of word2vec, the skip-gram model has significantly advanced the research of network embedding, such as the recent emergence of the DeepWalk, LINE, PTE, and node2vec approaches. In this work, we show that all of the aforementioned models with negative sampling can be unified into the matrix factorization framework with closed forms. Our analysis and proofs reveal that: (1) DeepWalk empirically produces a low-rank transformation of a network's normalized Laplacian matrix; (2) LINE, in theory, is a special case of DeepWalk when the size of vertices' context is set to one; (3) As an extension of LINE, PTE can be viewed as the joint factorization of multiple networks' Laplacians; (4) node2vec is factorizing a matrix related to the stationary distribution and transition probability tensor of a 2nd-order random walk. We further provide the theoretical connections between skip-gram based network embedding algorithms and the theory of graph Laplacian. Finally, we present the NetMF method as well as its approximation algorithm for computing network embedding. Our method offers significant improvements over DeepWalk and LINE for conventional network mining tasks. This work lays the theoretical foundation for skip-gram based network embedding methods, leading to a better understanding of latent network representation learning.

cs.SI↗

arXiv2017

Structural Diversity and Homophily: A Study Across More than One Hundred Big Networks

A widely recognized organizing principle of networks is structural homophily, which suggests that people with more common neighbors are more likely to connect with each other. However, what influence the diverse structures embedded in common neighbors have on link formation is much less well-understood. To explore this problem, we begin by characterizing the structural diversity of common neighborhoods. Using a collection of 120 large-scale networks, we demonstrate that the impact of the common neighborhood diversity on link existence can vary substantially across networks. We find that its positive effect on Facebook and negative effect on LinkedIn suggest different underlying networking needs in these networks. We also discover striking cases where diversity violates the principle of homophily---that is, where fewer mutual connections may lead to a higher tendency to link with each other. We then leverage structural diversity to develop a common neighborhood signature (CNS), which we apply to a large set of networks to uncover unique network superfamilies not discoverable by conventional methods. Our findings shed light on the pursuit to understand the ways in which network structures are organized and formed, pointing to potential advancement in designing graph generation models and recommender systems.

cs.SI↗

arXiv2017

Do the Young Live in a "Smaller World" Than the Old? Age-Specific Degrees of Separation in Human Communication

In this paper, we investigate the phenomenon of "age-specific small worlds" using data from a large-scale mobile communication network approximating interaction patterns at societal scale. Rather than asking whether two random individuals are separated by a small number of links, we ask whether individuals in specific age groups live in a small world in relation to individuals from other age groups. Our analysis shows that there is systematic variation in this age-relative small world effect. Young people live in the "smallest world," being separated from other young people and their parent's generation via a smaller number of intermediaries than older individuals. The oldest people live in the "least small world," being separated from their same age peers and their younger counterparts by a larger number of intermediaries. Variation in the small world effect is specific to age as a node attribute (being absent in the case of gender) and is consistently observed under several data robustness checks. The discovery of age-specific small worlds is consistent with well-known social mechanisms affecting the way age interacts with network connectivity and the relative prevalence of kin ties and non-kin ties observed in this network. This social pattern has significant implications for our understanding of generation-specific dynamics of information cascades, diffusion phenomena, and the spread of fads and fashions.

cs.SI↗

arXiv2017

A Century of Science: Globalization of Scientific Collaborations, Citations, and Innovations

Progress in science has advanced the development of human society across history, with dramatic revolutions shaped by information theory, genetic cloning, and artificial intelligence, among the many scientific achievements produced in the 20th century. However, the way that science advances itself is much less well-understood. In this work, we study the evolution of scientific development over the past century by presenting an anatomy of 89 million digitalized papers published between 1900 and 2015. We find that science has benefited from the shift from individual work to collaborative effort, with over 90% of the world-leading innovations generated by collaborations in this century, nearly four times higher than they were in the 1900s. We discover that rather than the frequent myopic- and self-referencing that was common in the early 20th century, modern scientists instead tend to look for literature further back and farther around. Finally, we also observe the globalization of scientific development from 1900 to 2015, including 25-fold and 7-fold increases in international collaborations and citations, respectively, as well as a dramatic decline in the dominant accumulation of citations by the US, the UK, and Germany, from ~95% to ~50% over the same period. Our discoveries are meant to serve as a starter for exploring the visionary ways in which science has developed throughout the past century, generating insight into and an impact upon the current scientific innovations and funding policies.

cs.DL↗

arXiv2016

Analysis of Link Formation, Persistence and Dissolution in NetSense Data

We study a unique behavioral network data set (based on periodic surveys and on electronic logs of dyadic contact via smartphones) collected at the University of Notre Dame.The participants are a sample of members of the entering class of freshmen in the fall of 2011 whose opinions on a wide variety of political and social issues and activities on campus were regularly recorded - at the beginning and end of each semester - for the first three years of their residence on campus. We create a communication activity network implied by call and text data, and a friendship network based on surveys. Both networks are limited to students participating in the NetSense surveys. We aim at finding student traits and activities on which agreements correlate well with formation and persistence of links while disagreements are highly correlated with non-existence or dissolution of links in the two social networks that we created. Using statistical analysis and machine learning, we observe several traits and activities displaying such correlations, thus being of potential use to predict social network evolution.

cs.SI↗

arXiv2016

Can Scientific Impact Be Predicted?

A widely used measure of scientific impact is citations. However, due to their heavy-tailed distribution, citations are fundamentally difficult to predict. Instead, to characterize scientific impact, we address two analogous questions asked by many scientific researchers: "How will my h-index evolve over time, and which of my previously or newly published papers will contribute to it?" To answer these questions, we perform two related tasks. First, we develop a model to predict authors' future h-indices based on their current scientific impact. Second, we examine the factors that drive papers---either previously or newly published---to increase their authors' predicted future h-indices. By leveraging relevant factors, we can predict an author's h-index in five years with an R2 value of 0.92 and whether a previously (newly) published paper will contribute to this future h-index with an F1 score of 0.99 (0.77). We find that topical authority and publication venue are crucial to these effective predictions, while topic popularity is surprisingly inconsequential. Further, we develop an online tool that allows users to generate informed h-index predictions. Our work demonstrates the predictability of scientific impact, and can help scholars to effectively leverage their position of "standing on the shoulders of giants."

cs.DL↗

arXiv2016

Influence Activation Model: A New Perspective in Social Influence Analysis and Social Network Evolution

What drives the propensity for the social network dynamics? Social influence is believed to drive both off-line and on-line human behavior, however it has not been considered as a driver of social network evolution. Our analysis suggest that, while the network structure affects the spread of influence in social networks, the network is in turn shaped by social influence activity (i.e., the process of social influence wherein one person's attitudes and behaviors affect another's). To that end, we develop a novel model of network evolution where the dynamics of network follow the mechanism of influence propagation, which are not captured by the existing network evolution models. Our experiments confirm the predictions of our model and demonstrate the important role that social influence can play in the process of network evolution. As well exploring the reason of social network evolution, different genres of social influence have been spotted having different effects on the network dynamics. These findings and methods are essential to both our understanding of the mechanisms that drive network evolution and our knowledge of the role of social influence in shaping the network structure.

physics.soc-ph↗

arXiv2016

Gender Differences in Communication Behaviors, Spatial Proximity Patterns, and Mobility Habits

The existence of gender differences in the structure and composition of social networks is a well established finding in the social and behavioral sciences, but researchers continue to debate whether structural, dispositional, or life course factors are the primary driver of these differences. In this paper we extend work on gender differences in social networks to patterns of interaction, propinquity, and connectivity captured via a social sensing platform comprised of an ensemble of individuals' phone calls, text messaging, face-to-face interactions, and traces of their mobility activities. We attempt to isolate dispositional from other factors by focusing on a relatively homogeneous population on a relatively closed setting at the same stage in the life course. Analysis across three different networks along with mobility data reveals systematic gender differences in terms of communicative, distributional, mobility, and spatial proximity tendencies. Most importantly, some patterns observed in the communication network (e.g. stronger same-gender preference for women) are found to be reversed in the spatial proximity context, with men displaying a greater tendency to spend time in a narrower (and thus more predictable) range of spaces with same-gender peers than women. These findings provide robust and novel evidence for the powerful effect of gender in structuring behavior across a wide range of communication and mobility behaviors, validating and extending recent work pointing to differences of socio-cultural and evolutionary origin in the styles of sociability and interaction characteristic of men and women.

physics.soc-ph↗

arXiv2015

Modeling the Interplay Between Individual Behavior and Network Distributions

It is well-known that many networks follow a power-law degree distribution; however, the factors that influence the formation of their distributions are still unclear. How can one model the connection between individual actions and network distributions? How can one explain the formation of group phenomena and their evolutionary patterns? In this paper, we propose a unified framework, M3D, to model human dynamics in social networks from three perspectives: macro, meso, and micro. At the micro-level, we seek to capture the way in which an individual user decides whether to perform an action. At the meso-level, we study how group behavior develops and evolves over time, based on individual actions. At the macro-level, we try to understand how network distributions such as power-law (or heavy-tailed phenomena) can be explained by group behavior. We provide theoretical analysis for the proposed framework, and discuss the connection of our framework with existing work. The framework offers a new, flexible way to explain the interplay between individual user actions and network distributions, and can benefit many applications. To model heavy-tailed distributions from partially observed individual actions and to predict the formation of group behaviors, we apply M3D to three different genres of networks: Tencent Weibo, Citation, and Flickr. We also use information-burst prediction as a particular application to quantitatively evaluate the predictive power of the proposed framework. Our results on the Weibo indicate that M3D's prediction performance exceeds that of several alternative methods by up to 30\%.

cs.SI↗

arXiv2015

Inferring Unusual Crowd Events From Mobile Phone Call Detail Records

The pervasiveness and availability of mobile phone data offer the opportunity of discovering usable knowledge about crowd behaviors in urban environments. Cities can leverage such knowledge in order to provide better services (e.g., public transport planning, optimized resource allocation) and safer cities. Call Detail Record (CDR) data represents a practical data source to detect and monitor unusual events considering the high level of mobile phone penetration, compared with GPS equipped and open devices. In this paper, we provide a methodology that is able to detect unusual events from CDR data that typically has low accuracy in terms of space and time resolution. Moreover, we introduce a concept of unusual event that involves a large amount of people who expose an unusual mobility behavior. Our careful consideration of the issues that come from coarse-grained CDR data ultimately leads to a completely general framework that can detect unusual crowd events from CDR data effectively and efficiently. Through extensive experiments on real-world CDR data for a large city in Africa, we demonstrate that our method can detect unusual events with 16% higher recall and over 10 times higher precision, compared to state-of-the-art methods. We implement a visual analytics prototype system to help end users analyze detected unusual crowd events to best suit different application scenarios. To the best of our knowledge, this is the first work on the detection of unusual events from CDR data with considerations of its temporal and spatial sparseness and distinction between user unusual activities and daily routines.

cs.SI↗

arXiv2015

Inferring Social Status and Rich Club Effects in Enterprise Communication Networks

Social status, defined as the relative rank or position that an individual holds in a social hierarchy, is known to be among the most important motivating forces in social behaviors. In this paper, we consider the notion of status from the perspective of a position or title held by a person in an enterprise. We study the intersection of social status and social networks in an enterprise. We study whether enterprise communication logs can help reveal how social interactions and individual status manifest themselves in social networks. To that end, we use two enterprise datasets with three communication channels --- voice call, short message, and email --- to demonstrate the social-behavioral differences among individuals with different status. We have several interesting findings and based on these findings we also develop a model to predict social status. On the individual level, high-status individuals are more likely to be spanned as structural holes by linking to people in parts of the enterprise networks that are otherwise not well connected to one another. On the community level, the principle of homophily, social balance and clique theory generally indicate a "rich club" maintained by high-status individuals, in the sense that this community is much more connected, balanced and dense. Our model can predict social status of individuals with 93% accuracy.

cs.SI↗