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 influences whom) typically very difficult. Furthermore, many applications, the underlying network over which diffusions propagations spread actually unobserved. We tackle these challenges by developing a method for tracing paths of influence through networks...

Large volumes of event data are becoming increasingly available in a wide variety applications, such as healthcare analytics, smart cities and social network analysis. The precise time interval or the exact distance between two events carries great deal information about dynamics underlying systems. These characteristics make fundamentally different from independently identically distributed time-series where space treated indexes rather than random variables. Marked temporal point processes...

Automated data-driven decision making systems are increasingly being used to assist, or even replace humans in many settings. These function by learning from historical decisions, often taken humans. In order maximize the utility of these (or, classifiers), their training involves minimizing errors misclassifications) over given data. However, it is quite possible that optimally trained classifier makes decisions for people belonging different social groups with misclassification rates...

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 or publish the information, observing individual transmissions (who infects whom, who influences whom) typically very difficult. Furthermore, many applications, underlying network over which diffusions propagations spread actually unobserved. We tackle these challenges by developing method for tracing paths of...

Algorithmic decision making systems are ubiquitous across a wide variety of online as well offline services. These rely on complex learning methods and vast amounts data to optimize the service functionality, satisfaction end user profitability. However, there is growing concern that these automated decisions can lead, even in absence intent, lack fairness, i.e., their outcomes disproportionately hurt (or, benefit) particular groups people sharing one or more sensitive attributes (e.g.,...

Time plays an essential role in the diffusion of information, influence and disease over networks. In many cases we only observe when a node copies makes decision or becomes infected -- but connectivity, transmission rates between nodes sources are unknown. Inferring underlying dynamics is outstanding interest since it enables forecasting, influencing retarding infections, broadly construed. To this end, model processes as discrete networks continuous temporal occurring at different rates....

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

Our work considers leveraging crowd signals for detecting fake news and is motivated by tools recently introduced Facebook that enable users to flag news. By aggregating users' flags, our goal select a small subset of every day, send them an expert (e.g., via third-party fact-checking organization), stop the spread identified as expert. The main objective minimize misinformation stopping propagation in network. It especially challenging achieve this it requires with high-confidence quickly...

Online social networking sites are experimenting with the following crowd-powered procedure to reduce spread of fake news and misinformation: whenever a user is exposed story through her feed, she can flag as misinformation and, if receives enough flags, it sent trusted third party for fact checking. If this identifies misinformation, marked disputed. However, given uncertain number exposures, high cost checking, trade-off between flags above mentioned requires careful reasoning smart...

We are witnessing an increasing use of data-driven predictive models to inform decisions in high-stakes situations, from lending and hiring university admissions. As have implications for individuals society, there is pressure on decision makers be transparent about their policies. At the same time, may knowledge, gained by transparency, invest effort strategically order maximize chances receiving a beneficial decision. Our goal find policies that optimal terms utility such strategic...

The combination of brain–computer interfaces (BCIs) with robot-assisted physical therapy constitutes a promising approach to neurorehabilitation patients severe hemiparetic syndromes caused by cerebrovascular brain damage (e.g. stroke) and other neurological conditions. In such scenario, key aspect is how reestablish the disrupted sensorimotor feedback loop. However, date it an open question artificially closing loop influences decoding performance BCI. this paper, we answer issue studying...

Deep generative models have been praised for their ability to learn smooth latent representation of images, text, and audio, which can then be used generate new, plausible data. However, current are unable work with molecular graphs due unique characteristics—their underlying structure is not Euclidean or grid-like, they remain isomorphic under permutation the nodes labels, come a different number edges. In this paper, we propose NeVAE, novel variational autoencoder graphs, whose encoder...

Abstract Time plays an essential role in the diffusion of information, influence, and disease over networks. In many cases we can only observe when a node is activated by contagion—when learns about piece makes decision, adopts new behavior, or becomes infected with disease. However, underlying network connectivity transmission rates between nodes are unknown. Inferring dynamics important because it leads to insights enables forecasting, as well influencing containing information...

Spaced repetition is a technique for efficient memorization which uses repeated review of content following schedule determined by spaced algorithm to improve long-term retention. However, current algorithms are simple rule-based heuristics with few hard-coded parameters. Here, we introduce flexible representation using the framework marked temporal point processes and then address design provable guarantees as an optimal control problem stochastic differential equations jumps. For two...

If a piece of information is released from media site, can it spread, in 1 month, to million web pages? This influence estimation problem very challenging since both the time-sensitive nature and issue scalability need be addressed simultaneously. In this paper, we propose randomized algorithm for continuous-time diffusion networks. Our estimate every node network with |V| nodes |E| edges an accuracy $\varepsilon$ using $n=O(1/\varepsilon^2)$ randomizations up logarithmic factors...

Information overload has become an ubiquitous problem in modern society. Social media users and microbloggers receive endless flow of information, often at a rate far higher than their cognitive abilities to process the information. In this paper, we conduct large scale quantitative study information evaluate its impact on dissemination Twitter social site. We model as processing systems that queue incoming according some policies, from unknown rates decide forward other users. show how...

Information diffusion in online social networks is affected by the underlying network topology, but it also has power to change it. Online users are constantly creating new links when exposed information sources, and turn these alternating way spreads. However, two highly intertwined stochastic processes, evolution, have been predominantly studied separately, ignoring their co-evolutionary dynamics. We propose a temporal point process model, COEVOLVE, for such joint dynamics, allowing...