面向關(guān)系利用的魯棒圖學(xué)習(xí)方法研究
發(fā)布時(shí)間：2019-07-04 15:36
【摘要】：關(guān)系在現(xiàn)實(shí)世界中無處不在。在機(jī)器學(xué)習(xí)研究領(lǐng)域,數(shù)據(jù)中有兩類關(guān)系不容忽視:1)樣本之間的關(guān)系;2)標(biāo)記之間的關(guān)系。大量研究結(jié)果表明,對這兩類關(guān)系的合理利用對提升訓(xùn)練模型的預(yù)測能力至關(guān)重要�；趫D的方法是關(guān)系利用的一類主流范型。這方面的代表性工作獲得了國際機(jī)器學(xué)習(xí)領(lǐng)域十年最佳論文獎(jiǎng)。經(jīng)過十余年的研究,基于圖的方法已取得了許多成果。然而,其學(xué)習(xí)性能嚴(yán)重依賴于圖的構(gòu)建�，F(xiàn)實(shí)任務(wù)中,圖構(gòu)建通常難以有效確定,使得學(xué)習(xí)性能的魯棒性不佳,有時(shí)還會(huì)出現(xiàn)性能的損害。本碩士論文圍繞提升關(guān)系利用的魯棒性這一重要問題展開研究,主要取得了以下創(chuàng)新成果:第一,針對樣本關(guān)系利用對圖構(gòu)建敏感的問題,提出了基于大間隔準(zhǔn)則的圖質(zhì)量判斷方法。該方法將魯棒樣本關(guān)系利用這一難題形式化為經(jīng)典半監(jiān)督支持向量機(jī)框架。優(yōu)化上給出高效的求解算法。實(shí)驗(yàn)結(jié)果表明,該方法顯著提升樣本關(guān)系利用的魯棒性,有效避免傳統(tǒng)方法會(huì)導(dǎo)致性能退化的現(xiàn)象。本論文還進(jìn)一步將大間隔準(zhǔn)則拓展用于帶噪樣本關(guān)系,提出了高效學(xué)習(xí)算法,有效防止帶噪樣本關(guān)系對性能的危害。第二,針對標(biāo)記關(guān)系利用對圖構(gòu)建敏感的問題,提出了基于分類器構(gòu)圈的標(biāo)記關(guān)系利用方法。該方法通過將分類器以圈形式構(gòu)建,克服了傳統(tǒng)學(xué)習(xí)方法在標(biāo)記關(guān)系利用中分類器次序?qū)π阅艿膰?yán)重影響。論文分析了該方法的時(shí)間復(fù)雜度與傳統(tǒng)方法相當(dāng),不顯著增加計(jì)算開銷。實(shí)驗(yàn)結(jié)果表明,該方法顯著提升標(biāo)記關(guān)系利用的魯棒性,有效避免傳統(tǒng)標(biāo)記關(guān)系利用方法會(huì)導(dǎo)致性能不佳的現(xiàn)象。
[Abstract]:Relationships are everywhere in the real world. In the field of machine learning, there are two kinds of relationships in the data that can not be ignored: 1) the relationship between samples and 2) the relationship between markers. A large number of research results show that the rational use of these two kinds of relations is very important to improve the prediction ability of the training model. The graph-based method is a kind of mainstream paradigm of relational utilization. The representative work in this area has won the ten-year best paper award in the field of international machine learning. After more than ten years of research, the graph-based method has made a lot of achievements. However, its learning performance depends heavily on the construction of graphs. In real tasks, graph construction is usually difficult to determine effectively, which makes the robustness of learning performance poor, and sometimes the performance damage. In this thesis, the important problem of improving the robustness of relational utilization is studied, and the following innovative results are obtained: first, aiming at the problem that the utilization of sample relationship is sensitive to graph construction, a graph quality judgment method based on large interval criterion is proposed. In this method, the robust sample relation is transformed into a classical semi-supervised support vector machine framework by using this problem. An efficient algorithm for solving the problem is given. The experimental results show that this method can significantly improve the robustness of sample relationship utilization and effectively avoid the phenomenon that the traditional method will lead to performance degradation. In this paper, the large interval criterion is further extended to the noisy sample relationship, and an efficient learning algorithm is proposed to effectively prevent the performance harm of the noisy sample relationship. Secondly, in order to solve the problem that mark relation utilization is sensitive to graph construction, a marker relation utilization method based on classification circle is proposed. By constructing the classifiers in the form of cycles, this method overcomes the serious influence of the order of classifiers on the performance of traditional learning methods in the utilization of tag relations. In this paper, the time complexity of this method is similar to that of the traditional method, and the computational overhead is not significantly increased. The experimental results show that this method can significantly improve the robustness of marker relationship utilization and effectively avoid the poor performance of traditional marker relationship utilization methods.
【學(xué)位授予單位】：南京大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP181
本文編號：2510022
[Abstract]:Relationships are everywhere in the real world. In the field of machine learning, there are two kinds of relationships in the data that can not be ignored: 1) the relationship between samples and 2) the relationship between markers. A large number of research results show that the rational use of these two kinds of relations is very important to improve the prediction ability of the training model. The graph-based method is a kind of mainstream paradigm of relational utilization. The representative work in this area has won the ten-year best paper award in the field of international machine learning. After more than ten years of research, the graph-based method has made a lot of achievements. However, its learning performance depends heavily on the construction of graphs. In real tasks, graph construction is usually difficult to determine effectively, which makes the robustness of learning performance poor, and sometimes the performance damage. In this thesis, the important problem of improving the robustness of relational utilization is studied, and the following innovative results are obtained: first, aiming at the problem that the utilization of sample relationship is sensitive to graph construction, a graph quality judgment method based on large interval criterion is proposed. In this method, the robust sample relation is transformed into a classical semi-supervised support vector machine framework by using this problem. An efficient algorithm for solving the problem is given. The experimental results show that this method can significantly improve the robustness of sample relationship utilization and effectively avoid the phenomenon that the traditional method will lead to performance degradation. In this paper, the large interval criterion is further extended to the noisy sample relationship, and an efficient learning algorithm is proposed to effectively prevent the performance harm of the noisy sample relationship. Secondly, in order to solve the problem that mark relation utilization is sensitive to graph construction, a marker relation utilization method based on classification circle is proposed. By constructing the classifiers in the form of cycles, this method overcomes the serious influence of the order of classifiers on the performance of traditional learning methods in the utilization of tag relations. In this paper, the time complexity of this method is similar to that of the traditional method, and the computational overhead is not significantly increased. The experimental results show that this method can significantly improve the robustness of marker relationship utilization and effectively avoid the poor performance of traditional marker relationship utilization methods.
【學(xué)位授予單位】：南京大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TP181
【相似文獻(xiàn)】
相關(guān)碩士學(xué)位論文 前1條
1 王少博;面向關(guān)系利用的魯棒圖學(xué)習(xí)方法研究[D];南京大學(xué);2017年
，本文編號：2510022
本文鏈接：http://sikaile.net/kejilunwen/zidonghuakongzhilunwen/2510022.html
最近更新
教材專著
