语义神经网络与自然语言深层语义的计算

Semantic Neural Networks and the Deep-Seated Semantic Computing of the Natural Language

本文依照语义神经网络、语义神经元及其上深层语义计算的形式化定义,基于深层语义计算需要分布式并行处理的特点,提出运用多线程技术来模拟实现神经元的同步并行计算。我们描述了如何利用多线程中单个线程实现神经元的独立自主运行,以及如何利用多线程中的同步机制实现神经元之间的消息传递。

In this paper, we present the formalized definition of Semantic Neural Network, the neurons in it and the deep semantic computing on it. Based on the characteristics of the deep-seated semantic computing which needs distributed parallel processing, we use a multithreading technique to simulate and actualize synchronized parallel processing. We describe how to use a single thread of multi-thread to actualize the independent running of neurons, and how to use the synchronized mechanism in multi-thread to actualize the transfer of information between neurons.