Short Description:

Blackouts, Congestions, Rumors, Diseases —— these different failures can be understood by one theoretical framework developed based on complex network theory. To mitigate these network failures, we need to understand mechanism behind the positive yet harmful feedback of failure propagation. Research of network reliability and resilience is aiming at exploring the cascading failure processes of different complex systems, evaluating the potential risks, predicting the possible scenarios, controlling the failure propagation, and finally building a 'smart' system.

Short Description:

The mysterious power modern artificial intelligence (AI) demonstrated in competing to beat top human players in chess and go, is also exploited to make urban life more convenient, less stressful and more enjoyable. Techniques including complex network analysis and artificial neural networks have been applied to trip planning, weather forcasting, autor-driving car designing, regional development, and medical diagnosis. The "AI in Urban Life" satellite conference of NetSCIX2018, organized by one of the largest non-government scientific associations in China, Swarma Club, aims to bridge the gap between industry and academia, and to create an opportunity for scientists across disciplines to exchange notes on their practices in applying AI techniques to serve for a better urban life.

Short Description:

The brain is a complex network. It perfectly extracts and integrates all kinds of information from various external and internal stimuli. To better explore the spatiotemporal pattern of the brain connectome. The human brain connectome is divided into three hierarchy: structural networks based on neuroanatomy; functional networks due to the dynamics of neuronal clusters, and effective networks that emphasize the causal interaction between nodes. Research of brain connectome is aiming to comprehend the dynamic evolution mechanism of the brain, understand the brain's advanced cognitive awareness, and ultimately return to ourselves.

Short Description:

Our world is networked: people are closer to each other through online social network services or mobile communication networks, while information is capable to be exchanged faster by World Wide Web or email networks. The social network is a treasure trove of user experiences and knowledge that presents great opportunities to understand the fundamental science of our world. In turn, many prediction tasks on nodes and edges have attracted considerable attention from both industry and academia. However, these tasks require careful effort in engineering features used by learning algorithms. While social network features require high computational resources and hard domain knowledge, it is critical to address the problem of learning network features automatically. Recent research in the broader field of network embedding, also known as representation learning for networks, has led to significant progress in automating prediction by learning the features themselves. The goal of network embedding is to project a network into a low-dimensional space, where each node can be presented as a single point in the learned latent space. However, many social network properties can not be captured by general network embedding algorithms. For instance, social networks are dynamic over time, while in most cases they are scale-free. This session aims to provide a forum for presenting the most recent advances in representation learning for social networks. We expect novel research on either frontier algorithms and models, or novel applications of network embedding on link prediction, fraud detection, network analysis, user modeling, and so on.