國立虎尾科技大學 |

Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition = = Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition =/
其他題名:	Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.
其他題名:	Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes
作者:	Tencer, Lukas.
面頁冊數:	1 online resource (223 pages)
附註:	Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.
Contained By:	Dissertation Abstracts International79-08B(E).
標題:	Computer science. -
電子資源:	click for full text (PQDT)
ISBN:	9780355798791

Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition = = Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.
Tencer, Lukas.

Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition =Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes - 1 online resource (223 pages)

Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.

Thesis (Ph.D.)--Ecole de Technologie Superieure (Canada), 2017.

Includes bibliographical references

The amount of data recorded and processed over recent years has increased exponentially. To create intelligent systems that can learn from this data, we need to be able to identify patterns hidden in the data itself, learn these pattern and predict future results based on our current observations. If we think about this system in the context of time, the data itself evolves and so does the nature of the classification problem. As more data become available, different classification algorithms are suitable for a particular setting. At the beginning of the learning cycle when we have a limited amount of data, online learning algorithms are more suitable. When truly large amounts of data become available, we need algorithms that can handle large amounts of data that might be only partially labeled as a result of the bottleneck in the learning pipeline from human labeling of the data.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9780355798791Subjects--Topical Terms:

573171
Computer science.
Index Terms--Genre/Form:

554714
Electronic books.

Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition = = Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.
LDR:08073ntm a2200433Ki 4500 001 920727
005 20181203094033.5
006 m o u
007 cr mn||||a|a||
008 190606s2017 xx obm 000 0 eng d
020 $a 9780355798791
035 $a (MiAaPQ)AAI10751464
035 $a (MiAaPQ)etsmh:10980
035 $a AAI10751464
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Tencer, Lukas. $3 1195600
245 1 0 $a Semi-Supervised Machine Learning Techniques for Classification of Evolving Data in Pattern Recognition = $b Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes.
246 3 1 $a Techniques Semi-Supervisees D'Apprentissage Machine Pour La Classification Des Donnees En Evolution En Reconnaissance De Formes
264 0 $c 2017
300 $a 1 online resource (223 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertation Abstracts International, Volume: 79-08(E), Section: B.
500 $a Adviser: Mohamed Cheriet.
502 $a Thesis (Ph.D.)--Ecole de Technologie Superieure (Canada), 2017.
504 $a Includes bibliographical references
520 $a The amount of data recorded and processed over recent years has increased exponentially. To create intelligent systems that can learn from this data, we need to be able to identify patterns hidden in the data itself, learn these pattern and predict future results based on our current observations. If we think about this system in the context of time, the data itself evolves and so does the nature of the classification problem. As more data become available, different classification algorithms are suitable for a particular setting. At the beginning of the learning cycle when we have a limited amount of data, online learning algorithms are more suitable. When truly large amounts of data become available, we need algorithms that can handle large amounts of data that might be only partially labeled as a result of the bottleneck in the learning pipeline from human labeling of the data.
520 $a An excellent example of evolving data is gesture recognition, and it is present throughout our work. We need a gesture recognition system to work fast and with very few examples at the beginning. Over time, we are able to collect more data and the system can improve. As the system evolves, the user expects it to work better and not to have to become involved when the classifier is unsure about decisions. This latter situation produces additional unlabeled data. Another example of an application is medical classification, where experts' time is a rare resource and the amount of received and labeled data disproportionately increases over time.
520 $a Although the process of data evolution is continuous, we identify three main discrete areas of contribution in different scenarios. When the system is very new and not enough data are available, online learning is used to learn after every single example and to capture the knowledge very fast. With increasing amounts of data, offline learning techniques are applicable. Once the amount of data is overwhelming and the teacher cannot provide labels for all the data, we have another setup that combines labeled and unlabeled data. These three setups define our areas of contribution; and our techniques contribute in each of them with applications to pattern recognition scenarios, such as gesture recognition and sketch recognition.
520 $a An online learning setup significantly restricts the range of techniques that can be used. In our case, the selected baseline technique is the Evolving TS-Fuzzy Model. The semi-supervised aspect we use is a relation between rules created by this model. Specifically, we propose a transductive similarity model that utilizes the relationship between generated rules based on their decisions about a query sample during the inference time. The activation of each of these rules is adjusted according to the transductive similarity, and the new decision is obtained using the adjusted activation. We also propose several new variations to the transductive similarity itself.
520 $a These three works are connected by gradually relaxing the constraints on the learning setting in which we operate. Although our main motivation behind the development was to increase performance in various real-world tasks (gesture recognition, sketch recognition), we formulated our work as general methods in such a way that they can be used outside a specific application setup, the only restriction being that the underlying data evolve over time. Each of these methods can successfully exist on its own. The best setting in which they can be used is a learning problem where the data evolve over time and it is possible to discretize the evolutionary process. (Abstract shortened by ProQuest.).
520 $a La quantite de donnees enregistrees et traitees au cours des dernieres annees a augmente de facon exponentielle. Pour creer des systemes intelligents qui peuvent apprendre de ces donnees, nous devons etre en mesure d'identifier les modeles caches dans les donnees elles-memes, apprendre ces modeles et predire les resultats futurs sur la base de nos observations actuelles. Si nous pensons a ce systeme dans un contexte temporel, les donnees elles-memes evoluent, tout comme la nature du probleme de classification. Lorsque plus de donnees deviennent disponibles, differents algorithmes de classification sont adaptes a un contexte particulier. Au debut de la phase d'apprentissage lorsque nous disposons d'une quantite limitee de donnees d'entrainement, les algorithmes d'apprentissage en ligne sont plus appropries. Lorsque de grandes quantites de donnees deviennent disponibles, nous avons besoin d'algorithmes qui peuvent traiter de grandes quantites de donnees partiellement etiquetees dues a la limitation d'etiquetage manuel.
520 $a Un exemple typique ou les donnees evoluent est la reconnaissance de geste. Ce dernier exemple est present tout au long de notre travail. Nous avons besoin des systemes de reconnaissance des gestes pour fonctionner rapidement et avec tres peu d'echantillons au debut. Au cours du temps, nous sommes en mesure de collecter plus de donnees pour que la performance du systeme s'ameliore. A mesure que le systeme evolue, l'utilisateur s'attend a ce qu'il fonctionne mieux et qu'il n'ait plus besoin de s'impliquer lorsque le classificateur est incertain quant aux decisions. Dans cette derniere situation des donnees supplementaires non etiquetees sont alors produites. Un autre exemple typique d'une application est la classification de donnees medicales, ou le temps des experts (cliniciens, chirurgiens) est une ressource rare et la quantite de donnees recues et etiquetees augmente de facon desequilibree au cours du temps. Bien que le processus de l'evolution des donnees soit continu, nous pouvons identifier trois contributions dans differents scenarios. Lorsque le systeme est nouveau avec peu de donnees, l'apprentissage en ligne est utilise pour apprendre apres chaque echantillon et capturer les connaissances tres rapidement. Avec l'augmentation de quantites de donnees, les techniques d'apprentissage hors ligne deviennent davanatge applicables. Une fois que la quantite de donnees est massive et que le processus d'etiquetage ne couvre pas toutes les donnees, nous avons une autre configuration qui combine les donnees etiquetees et celles non etiquetees. Ces trois configurations definissent nos axes de contributions avec comme applications la reconnaissance des gestes et la reconnaissance de croquis en ligne.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Computer engineering. $3 569006
650 4 $a Artificial intelligence. $3 559380
655 7 $a Electronic books. $2 local $3 554714
690 $a 0984
690 $a 0464
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Ecole de Technologie Superieure (Canada). $b Departement de genie de la production automatisee. $3 1179220
773 0 $t Dissertation Abstracts International $g 79-08B(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10751464 $z click for full text (PQDT)