项目名称: 基于非线性搜索的量子神经计算模型研究
项目编号: No.61065002
项目类型: 地区科学基金项目
立项/批准年度: 2011
项目学科: 金属学与金属工艺
项目作者: 周日贵
作者单位: 华东交通大学
项目金额: 11万元
中文摘要: 针对传统神经计算记忆容量有限、数据处理速度慢及易发生灾变性失忆等缺点,本项目采用量子计算与神经计算相结合的新方法,充分发挥量子计算中量子线性叠加特性的巨大计算优势,应用存储矩阵元素基于概率分布的权值矩阵确定方法,建立了多模式并存的量子神经计算模型,使得该模型的存储容量或记忆容量提高到神经元个数的2的n次方倍,较传统神经计算模型有指数级的提高;最后用人脸图像对模型及算法进行了验证。本项目对模式识别和超大容量图像处理等诸多领域的研究将提供重要的理论支持并产生潜在的实用价值。
中文关键词: 量子神经计算;概率分布;时间复杂度
英文摘要: Due to limited memory capacity,slowly data processing and Catastrophic forgetting of traditional neural computing,we adopt new method of combining quantum computing and neural computation,give full play to tremendous computing superiority of quantum linear superposition,apply elements of the storage matrix that distributed in a probability way to determine the weight matrix and build model of quantum neural computation that can deal with multi-pattern.The storage capacity of this model is increased by a factor of 2N, where N is the number of neurons,comparing to the conventional neural network.In the last,we test and verify the medel and algorithm using people face images.This project will provide important theory supporting to the many fields such as pattern recognition and super high-capacity image processing and produce potential practical value.
英文关键词: quantum neural computing;probability distribution;time complexity