We present a novel column generation based boosting method for multi-class classification. Our multi-class boosting is formulated in a single optimization problem as in [1, 2]. Different from most existing multi-class boosting methods, which use the same set of weak learners for all the classes, we train class specified weak learners (i.e., each class has a different set of weak learners). We s...

Scaling discrete AdaBoost to handle real-valued weak hypotheses has often been done under the auspices of convex optimization, but little is generally known from the original boosting model standpoint. We introduce a novel generalization of discrete AdaBoost which departs from this mainstream of algorithms. From the theoretical standpoint, it formally displays the original boosting property, as...

A key problem in visual tracking is how to handle the ambiguity in decision to locate the object effectively using the target appearance model with online update. We address this problem by incorporating sequential clustering and ensemble methods into the tracking system. In this paper, clustering is used for mining the potential historical structure in the parameter space and feature space. Th...

Boosting is one of the most significant development in machine learning areas in recent years. Although boosting has already achieved great success in practical applications, its internal mechanism has not been entirely understood. In this paper, we present a new perspective to design boosting algorithms: extracting independent weak rules. A boosting algorithm can be divided into two parts, an ...

Open learner models (OLM) are learner models which are accessible to the learner, allowing them to view, and sometimes modify, their model. This openness may raise questions of learner trust in their learner model: if users do not agree with, or trust the information they can see about themselves, their trust in the interaction will likely be reduced. Using a Wizard-of-Oz approach, we consider ...

Major objective of the web-based learning is to design customized courseware taking into account of the input knowledge level and the specific objective of each individual learner. In order to improve the design of adaptive dynamic courseware to meet the requirement of individual learner we have already presented concept of Learner Quanta (LQ) & Learner Quanta Cloud (LQC) [9]. In this paper, co...

Using an eye-tracker, this paper investigates the information that learners visually attend to in their open learner model, and the degree to which this is related to the method of displaying the model to the learner. Participants were fourteen final year undergraduate students using six views of their learner model data. Results suggest some views of the learner model information may be more l...