Asymptotic Properties of a Dynamic Neural System with Asymmetric Connection Weights

In this paper, based on new Lyapunov function, the asymptotic properties of the dynamic neural system with asymmetric connection weights are investigated. Since the dynamic neural system with asymmetric connection weights is more general than that with symmetric ones, the new results are significant in both theory and applications. Specially the new result can cover the asymptotic stability results of linear systems as special cases.

The Asymmetric Connection of SST in the Tasman Sea with Respect to the Opposite Phases of ENSO in Austral Summer

This study uses linear regression and composite analyses to identify a pronounced asymmetric connection of sea surface temperature(SST)in the Tasman Sea with the two opposite phases of El Ni?o-Southern Oscillation(ENSO)during austral summer.In El Ni?o years,the SST anomalies(SSTAs)in the Tasman Sea exhibit a dipolar pattern with weak warmth in the northwest and modest cooling in the southeast,while during La Ni?a years the SSTAs exhibit a basin-scale warmth with greater amplitude.Investigations into the underlying mechanism suggest that this asymmetry arises from a mechanism related to oceanic heat transport,specifically the anomalous Ekman meridional heat transport induced by the zonal wind stress anomalies,rather than the surface heat fluxes on the air-sea interface.Further analysis reveals that the asymmetry of oceanic heat transport between El Ni?o and La Ni?a years is driven by the asymmetric atmospheric circulation over the Tasman Sea stimulated by the asymmetric diabatic heating in the tropical Pacific between the two opposite ENSO phases.

Stability of a Class of Neural Networks with Asymmetric Connection Weights

This paper derives some sufficient conditions for exponential stability for the equilibrium point by dividing the state variables of the system according to the characters of the neural networks. The new conditions are described by some blocks of the interconnection matrix. An example is given to demonstrate the effectiveness of the proposed theory.

Analyzing Asymmetric City Connectivity by Toponym on Social Media in China

The connectedness between cities has become one of the most widely discussed topics in urban and regional research in the mobile and big data era. One problem identified is the asymmetric city connectivity, partially due to data availability. We present a data-driven approach based on location and toponym(place name) extracted from social media data, to assess the asymmetric connectivity between cities. The assumption is that a higher frequency of occurrences of the name of city i in posts located in city j would imply that the city i is more influential than other cities upon city j. In addition, we’ve developed a group of measurements such as the relatedness index, impact index, link strength index, dependence index, and structure similar index to characterize such interactions. This framework of connectivity measurements can also be used to support smart planning taking into account the evolving interplay among cities. The space-time structure of urban systems in China is examined as the case study.

Dynamical parton distributions from DGLAP equations with nonlinear corrections

Determination of proton parton distribution functions is presented under the dynamical parton model assumption by applying DGLAP equations with GLR-MQ-ZRS corrections.We provide two data sets,referred to as IMParton16,which are from two different nonperturbative inputs.One is the naive input of three valence quarks and the other is the input of three valence quarks with flavor-asymmetric sea components.Basically,both data sets are compatible with the experimental measurements at high scale（Q^2〉 2 GeV^2）.Furthermore,our analysis shows that the input with flavor-asymmetric sea components better reproduces the structure functions at high Q^2.Generally,the parton distribution functions obtained,especially the gluon distribution function,are good options for inputs to simulations of high energy scattering processes.The analysis is performed under the fixed-flavor number scheme for nf = 3,4,5.Both data sets start from very low scales,around 0.07 GeV^2,where the nonperturbative input is directly connected to the simple picture of the quark model.These results may shed some lights on the origin of the parton distributions observed at high Q^2.