Chi Ho Yeung

Networking—a statistical physics perspective

Networking encompasses a variety of tasks related to the communication of information on networks; it has a substantial economic and societal impact on a broad range of areas including transportation systems, wired and wireless communications and a range of Internet applications. As transportation and communication networks become increasingly more complex, the ever increasing demand for congestion control, higher traffic capacity, quality of service, robustness and reduced energy consumption requires new tools and methods to meet these conflicting requirements. The new methodology should serve for gaining better understanding of the properties of networking systems at the macroscopic level, as well as for the development of new principled optimization and management algorithms at the microscopic level. Methods of statistical physics seem best placed to provide new approaches as they have been developed specifically to deal with nonlinear large-scale systems. This review aims at presenting an overview of tools and methods that have been developed within the statistical physics community and that can be readily applied to address the emerging problems in networking. These include diffusion processes, methods from disordered systems and polymer physics, probabilistic inference, which have direct relevance to network routing, file and frequency distribution, the exploration of network structures and vulnerability, and various other practical networking applications.

The reinforcing influence of recommendations on global diversification

Recommender systems are promising ways to filter the abundant information in modern society. Their algorithms help individuals to explore decent items, but it is unclear how they distribute popularity among items. In this paper, we simulate successive recommendations and measure their influence on the dispersion of item popularity by Gini coefficient. Our result indicates that local diffusion and collaborative filtering reinforce the popularity of hot items, widening the popularity dispersion. On the other hand, the heat conduction algorithm increases the popularity of the niche items and generates smaller dispersion of item popularity. Simulations are compared to mean-field predictions. Our results suggest that recommender systems have reinforcing influence on global diversification. Finally, the study of the hybrid method of mass diffusion and heat conduction reveals that the influence of recommender systems is actually controllable.

From the physics of interacting polymers to optimizing routes on the London Underground

Optimizing paths on networks is crucial for many applications, ranging from subway traffic to Internet communication. Because global path optimization that takes account of all path choices simultaneously is computationally hard, most existing routing algorithms optimize paths individually, thus providing suboptimal solutions. We use the physics of interacting polymers and disordered systems to analyze macroscopic properties of generic path optimization problems and derive a simple, principled, generic, and distributed routing algorithm capable of considering all individual path choices simultaneously. We demonstrate the efficacy of the algorithm by applying it to: (i) random graphs resembling Internet overlay networks, (ii) travel on the London Underground network based on Oyster card data, and (iii) the global airport network. Analytically derived macroscopic properties give rise to insightful new routing phenomena, including phase transitions and scaling laws, that facilitate better understanding of the appropriate operational regimes and their limitations, which are difficult to obtain otherwise.

Heterogenous human dynamics in intra- and inter-day time scales

In this paper, we study two large data sets containing the information of two different human behaviors: blog-posting and wiki-revising. In both cases, the interevent time distributions decay as power laws at both individual and population level. Unlike previous studies, we put emphasis on time scales and obtain heterogeneous decay exponents in the intra- and inter-day range for the same dataset. Moreover, we observe opposite trend of exponents in relation to individual Activity. Further investigations show that the presence of intra-day activities mask the correlation between consecutive inter-day activities and lead to an underestimate of Memory, which explain the contradicting results in recent empirical studies. Removal of data in intra-day range reveals the high values of Memory and lead us to convergent results between wiki-revising and blog-posting.

Recovery of infrastructure networks after localised attacks

The stability of infrastructure network is always a critical issue studied by researchers in different fields. A lot of works have been devoted to reveal the robustness of the infrastructure networks against random and malicious attacks. However, real attack scenarios such as earthquakes and typhoons are instead localised attacks which are investigated only recently. Unlike previous studies, we examine in this paper the resilience of infrastructure networks by focusing on the recovery process from localised attacks. We introduce various preferential repair strategies and found that they facilitate and improve network recovery compared to that of random repairs, especially when population size is uneven at different locations. Moreover, our strategic repair methods show similar effectiveness as the greedy repair. The validations are conducted on simulated networks, and on real networks with real disasters. Our method is meaningful in practice as it can largely enhance network resilience and contribute to network risk reduction.

Competition for shortest paths on sparse graphs

Optimal paths connecting randomly selected network nodes and fixed routers are studied analytically in the presence of a nonlinear overlap cost that penalizes congestion. Routing becomes more difficult as the number of selected nodes increases and exhibits ergodicity breaking in the case of multiple routers. The ground state of such systems reveals nonmonotonic complex behaviors in average path length and algorithmic convergence, depending on the network topology, and densities of communicating nodes and routers. A distributed linearly scalable routing algorithm is also devised.

Enhancing synchronization by directionality in complex networks

We propose a method called the residual edge-betweenness gradient (REBG) to enhance the synchronizability of networks by assigning the link direction while keeping the topology and link weights unchanged. Direction assignment has been shown to improve the synchronizability of undirected networks in general, but we find that in some cases incommunicable components emerge and networks fail to synchronize. We show that the REBG method improves the residual degree gradient (RDG) method by effectively avoiding the synchronization failure. Further experiments show that the REBG method enhances the synchronizability in networks with a community structure compared with the RDG method.

Heterogenous scaling in the inter-event time of on-line bookmarking

In this paper, we study the statistical properties of bookmarking behaviors in Delicious.com. We find that the inter-event time (τ) distributions of bookmarking decay in a power-like manner as τ increases at both individual and population levels. Remarkably, we observe a significant change in the exponent when the inter-event time increases from the intra-day range to the inter-day range. In addition, the dependence of the exponent on individual activity is found to be different in the two ranges. Instead of monotonically increasing with activity, the inter-day exponent peaks around 3. These results suggest that the mechanisms driving human actions are different in the intra-day and inter-day ranges. We further show that the global distributions of less active users are closer to an exponential distribution than those of more active users. Moreover, a universal behavior in the inter-day range is observed by considering the rescaled variable τ/〈τ〉. Finally, the possible causes of these phenomena are discussed.

How to quantify the influence of correlations on investment diversification

When assets are correlated, benefits of investment diversification are reduced. To measure the influence of correlations on investment performance, a new quantity - the effective portfolio size - is proposed and investigated in both artificial and real situations. We show that in most cases, the effective portfolio size is much smaller than the actual number of assets in the portfolio and that it lowers even further during financial crises.

Tracing the evolution of physics on the backbone of citation networks

Many innovations are inspired by past ideas in a nontrivial way. Tracing these origins and identifying scientific branches is crucial for research inspirations. In this paper, we use citation relations to identify the descendant chart, i.e., the family tree of research papers. Unlike other spanning trees that focus on cost or distance minimization, we make use of the nature of citations and identify the most important parent for each publication, leading to a treelike backbone of the citation network. Measures are introduced to validate the backbone as the descendant chart. We show that citation backbones can well characterize the hierarchical and fractal structure of scientific development, and lead to an accurate classification of fields and subfields.