Stability has been explored to study the performance of learning algorithms in recent years and it has been shown that stability is sufficient for generalization and is sufficient and necessary for consistency of ERM in the general learning setting. Previous studies showed that AdaBoost has almost-everywhere uniform stability if the base learner has L1 stability. The L1 stability, however, is t...

No learner is generally better than another learner. If a learner performs better than another learner on some learning situations, then the first learner usually performs worse than the second learner on other situations. In other words, no single learning algorithm can perform well and uniformly outperform other algorithms over all learning or data mining tasks. There is an increasing number ...

This paper presents a brief review of early learner modelling in Intelligent Tutoring Systems focusing particularly on procedural knowledge representations vs. declarative knowledge representations. It then tracks the paradigm shift from traditional learner modelling, which emphasises knowledge representation, to distributed learner modelling, which focuses on the modelling process. Learner mod...

There has been almost no work on phrase structure annotation and parsing specially designed for learner English despite the fact that they are useful for representing the structural characteristics of learner English. To address this problem, in this paper, we first propose a phrase structure annotation scheme for learner English and annotate two different learner corpora using it. Second, we s...

Learner Modeling is a dynamic process. Creating and maintaining models of learners by an interactive system is not a precise task, and it requires guesses and abductive generation of explanation of the learners’ actions. In addition, learners change their behavior while interacting with the system. Therefore, these systems need some mechanisms to maintain the rationality of the learner models c...

To advance our understanding of learning in massive open online courses (MOOCs), we need to understand how learners interact with course resources. Prior explorations of learner interactions with MOOC materials have often described these interactions through stereotypes, which does not account for the full spectrum of potential learner activities. A focus on stereotypes also limits our ability ...

In the framework of kit-build concept map, (1) a goal map is prepared by a teacher, (2) parts of the map is generated by decomposing the goal map, and then, (3) a learner makes a map (a learner map) by combining the parts. The learner map is diagnosed by comparing with the goal map. The differences between the learner map and the goal map should be resolved by the learner to complete the learne...

We introduce a formal model of teaching in which the teacher is tailored to a particular learner, yet the teaching protocol is designed so that no collusion is possible. Not surprisingly, such a model remedies the non-intuitive aspects of otehr models in which the teacher must successfully teach any consistent learner. We prove that any class that can be exactly identified by a deterministic po...

The first two papers address issues surrounding learner inspection of peer models, and the availability of learner models to course instructors. Bull, Mabbott and Abu Issa present the UMPTEEN system which allows students to open their learner models to others in named and/or anonymous form. The CosyQTI system presented by Lazarinis and Retalis provides anonymous access to models of peers and re...

This paper is a comprehensive research review on the learner control strategies in software tutoring systems. With the application of more computer techniques in education and the involvement of more adults in software tutoring systems, the learner control strategy has become more appreciated than tutor control or program control. In this paper, the efficiency and necessity of learner control i...