APPROXIMATING TRUST FUNCTIONS FROM MODELS IN NEURON NETWORK AND LIE ALGEBRA

Abstract

This paper is to describe a tensor representation of features needed

for trust computing in complex networks. Then we propose a novel hy-

brid framework that functionally approximates trust functions defined

over multi-dimensional trust tensors using both neural networks and Lie

algebra-based mappings. The approximation capability of neural net-

works in this context and utilize Lie group structures are to capture

structural symmetries and behaviors in trust propagation.