Multiple Incremental Decremental Learning of Support Vector Machines

Karasuyama, Masayuki; Takeuchi, Ichiro

doi:10.1109/tnn.2010.2048039

Cited by 110 publications

(75 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, the incremental classifier methods target on updating the prediction boundary with the learned model parameters and new samples. Exemplars include the incremental support vector machines (ISVM) [24] and the online sequential forward neural network [25]. In addition, several attempts have been made to absorb advantages from both of the two categories of methods.…”

Section: Related Workmentioning

confidence: 99%

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Lin

Wang

Meng

et al. 2018

IEEE Trans. Pattern Anal. Mach. Intell.

151

View full text Add to dashboard Cite

Abstract-This paper aims to develop a novel cost-effective framework for face identification, which progressively maintains a batch of classifiers with the increasing face images of different individuals. By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into training under weak expert recertification. We first initialize the classifier using a few annotated samples for each individual, and extract image features using the convolutional neural nets. Then, a number of candidates are selected from the unannotated samples for classifier updating, in which we apply the current classifiers ranking the samples by the prediction confidence. In particular, our approach utilizes the high-confidence and low-confidence samples in the selfpaced and the active user-query way, respectively. The neural nets are later fine-tuned based on the updated classifiers. Such heuristic implementation is formulated as solving a concise active SPL optimization problem, which also advances the SPL development by supplementing a rational dynamic curriculum constraint. The new model finely accords with the "instructorstudent-collaborative" learning mode in human education. The advantages of this proposed framework are two-folds: i) The required number of annotated samples is significantly decreased while the comparable performance is guaranteed. A dramatic reduction of user effort is also achieved over other state-of-theart active learning techniques. ii) The mixture of SPL and AL effectively improves not only the classifier accuracy compared to existing AL/SPL methods but also the robustness against noisy data. We evaluate our framework on two challenging datasets, which include hundreds of persons under diverse conditions, and demonstrate very promising results. Please find the code of this project at: http://hcp.sysu.edu.cn/projects/aspl/

show abstract

Section: Related Workmentioning

confidence: 99%

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Lin

Wang

Meng

et al. 2018

IEEE Trans. Pattern Anal. Mach. Intell.

151

View full text Add to dashboard Cite

show abstract

“…The regularization parameter C balances the importance of error and margin [9]. In practice, the dual problem of (4) is solved to obtain a SVM classifier [10], [11].…”

Section: Preliminary Knowledgementioning

confidence: 99%

Distributed incremental wLPSVM learning

Zhu

Ban

Ikeda

et al. 2016

2016 IEEE Symposium Series on Computational Intelligence (SSCI)

View full text Add to dashboard Cite

Abstract-Weighted linear proximal support vector machine (wLPSVM) is known as an efficient binary classification algorithm with good accuracy and class-imbalance robustness. In this work, original batch wLPSVM is facilitated with distributed incremental learning capability, which allows simultaneously learning from multiple streaming data sources that are geographically distributed. In our approach, incremental and distributed learning are solved as a merging problem at the same time.A new wLPSVM expression is derived. In the new expression, knowledge from samples are presented as a set of class-wised core matrices, and merging knowledge from two subsets of data can be simply accomplished by matrix addition. With the new expression, we are able to conduct incremental and distributed learning at the same time via merging knowledge from multiple incremental stages and multiple data sources.

show abstract

“…The regularization parameter C balances the importance of error and margin (Yamasaki & Ikeda, 2005). In practice, the dual problem of (6) is solved to obtain a SVM classifier (Cauwenberghs & Poggio, 2000;Karasuyama & Takeuchi, 2010). LPSVM (Fung & Mangasarian, 2001b) models a binary classification as a regularized least square problem, which simplifies the above SVM optimization and results in an extremely efficient training algorithm.…”

Section: Batch Lpsvmmentioning

confidence: 99%

Dynamic class imbalance learning for incremental LPSVM

et al. 2013

View full text Add to dashboard Cite

a b s t r a c tLinear Proximal Support Vector Machines (LPSVMs), like decision trees, classic SVM, etc. are originally not equipped to handle drifting data streams that exhibit high and varying degrees of class imbalance. For online classification of data streams with imbalanced class distribution, we propose a dynamic class imbalance learning (DCIL) approach to incremental LPSVM (IncLPSVM) modeling. In doing so, we simplify a computationally non-renewable weighted LPSVM to several core matrices multiplying two simple weight coefficients. When data addition and/or retirement occurs, the proposed DCIL-IncLPSVM 1 accommodates newly presented class imbalance by a simple matrix and coefficient updating, meanwhile ensures no discriminative information lost throughout the learning process. Experiments on benchmark datasets indicate that the proposed DCIL-IncLPSVM outperforms classic IncSVM and IncLPSVM in terms of F -measure and G-mean metrics. Moreover, our application to online face membership authentication shows that the proposed DCIL-IncLPSVM remains effective in the presence of highly dynamic class imbalance, which usually poses serious problems to previous approaches.

show abstract

Multiple Incremental Decremental Learning of Support Vector Machines

Cited by 110 publications

References 16 publications

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Distributed incremental wLPSVM learning

Dynamic class imbalance learning for incremental LPSVM

Contact Info

Product

Resources

About