Deep-seated Semantic Computing Based On Semantic Neural Network

Semantic Neural Network is a new methodology for Natural Language Understanding that Combines symbolism and connectionism. It breaks the traditional linear understanding model and simulates a mechanism for language processor of the human brain, it is considered that the Natural Language Understanding is the process that the linguistic labels activate the corresponding neuron in the human brain, and set up or activate the Semantic Neural Network. Deep-seated Semantic Computing of Semantic Neural Network is that continue to use Deep-seated Semantic knowledge embed in neuron, and through each other's communication and inside computing to finish this processing. The result of Deep-seated Semantic Computing in whole Semantic Neural Network is a result of nature language understanding. If we can achieve this idea, its significance will tell its own tale. At the present we have made a little progress in this research. [1] has provided the idea and its models, [28] has an in-depth study of the global model of Semantic Neural Network, the design of the neuron and the analysis of the Chinese Surface Semantics, and through the analysis of some examples, it is apparent that the feasibility of carrying on the understanding of the surface semantics with Semantic Neural Network. On the basis of this work, we continue to carry on the research of Deep-seated Semantic Computing and really realize the understanding of the nature language with the Semantic Neural Network. In this paper, we perfect the designs of the model and structure of the neuron in the Semantic Neural Network, describe the structure course of the Chinese Surface Semantics Neural Network, and Recommend the Deep-seated semantics of the neuron to define and formalization description. Variety on the basis of the Deep-seated semantic knowledge, we classify to Deep-seated Semantic Computing, and different Deep-seated semantic knowledge use different computing technologies to deal with. So we design a fact base and a rule base for each neuron, the fact base is used to store the Deep-seated semantic knowledge of the neuron, it will activate a corresponding Deep-seated semantic computing method each time while facts activate a rule. It is a distributed computational process that runs side by side that the Deep-seated semantics of the Semantic Neural Network is computed, we adopt multithreading technology to imitate independence and the initiative of the neuron and distributed parallel computation. This text did some experiment examples finally, the result shows that it is feasible to realize nature language understanding with the Semantics Neural Network.