二進(jìn)神經(jīng)網(wǎng)絡(luò)的模式匹配學(xué)習(xí)
The pattern match learning of binary neural networks
-
摘要: 二進(jìn)神經(jīng)網(wǎng)絡(luò)的知識(shí)提取需要了解每個(gè)神經(jīng)元的邏輯意義。一般來(lái)說(shuō),對(duì)二進(jìn)神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)結(jié)果的分析是困難的。該文提出了一種基于線性可分結(jié)構(gòu)系結(jié)構(gòu)分析的學(xué)習(xí)算法,采用這種方法對(duì)布爾空間的樣本集合進(jìn)行學(xué)習(xí),得到的二進(jìn)神經(jīng)網(wǎng)絡(luò)隱層神經(jīng)元都?xì)w屬于一類或幾類線性可分結(jié)構(gòu)系,只要這幾類線性可分結(jié)構(gòu)系的邏輯意義是清晰的,就可以分析整個(gè)學(xué)習(xí)結(jié)果的知識(shí)內(nèi)涵。
-
關(guān)鍵詞:
- 二進(jìn)神經(jīng)網(wǎng)絡(luò); 線性可分; 模式匹配
Abstract: It is necessary to know the logical meaning of every binary neuron when extracting knowledge from a binary neural network. Generally, it is difficult, to analyze learning results of a learning algorithm for binary neural networks. Ln this paper, a new learning method is presented which is based on analyzing a set of linear separable structures. The most important benefit of this method is all binary neurons belong to one or more types of linear separable structure sets. If those linear separable structure sets have clear logical meaning, the whole knowledge of binary neural networks can be dug out. -
陸陽(yáng),韓江洪,高雋,魏臻,二進(jìn)神經(jīng)網(wǎng)絡(luò)中漢明球的邏輯意義及一般判別方法,計(jì)算機(jī)研究與發(fā)展,2002,39(1),79-86 ［2］J.H.Kim,S.Park,The geometrical learning of binary neural networks,IEEE Trans.on Neural Networks,1995,6(1),237-247. ［3］朱大銘,馬紹漢,二進(jìn)制神經(jīng)網(wǎng)絡(luò)分類問(wèn)題的幾何學(xué)習(xí)算法,軟件學(xué)報(bào),1997,8(8),622-629. ［4］馬曉敏,楊義先,章照止,一種新的閾函數(shù)的分析框架及有關(guān)結(jié)論,計(jì)算機(jī)學(xué)報(bào),2000,23(3),225-230. ［5］陸陽(yáng),韓江洪,張維勇,二進(jìn)神經(jīng)網(wǎng)絡(luò)邏輯關(guān)系判據(jù)及等價(jià)性規(guī)則提取,模式識(shí)別與人工智能,2001,14(2),171-176. ［6］陸陽(yáng),韓江洪,高雋,二進(jìn)神經(jīng)網(wǎng)絡(luò)隱元數(shù)目最小上界研究,模式識(shí)別與人工智能,2000,13(3),254-257. ［7］馬曉敏,楊義先,章照止,二進(jìn)神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)算法研究,計(jì)算機(jī)學(xué)報(bào),1999,22(9),931-935. -
計(jì)量
- 文章訪問(wèn)數(shù): 2428
- HTML全文瀏覽量: 101
- PDF下載量: 518
- 被引次數(shù): 0
下載:
