Conference Paper

Building Pedagogical Models by Formal Concept Analysis

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

The Pedagogical Model is one of the main components of an Intelligent Tutoring System. It is exploited to select a suitable action (e.g., feedback, hint) that the intelligent tutor provides to the learner in order to react to her interaction with the system. Such selection depends on the implemented pedagogical strategy and, typically, takes care of several aspects such as correctness and delay of the learner’s response, learner’s profile, context and so on. The main idea of this paper is to exploit Formal Concept Analysis to automatically learn pedagogical models from data representing human tutoring behaviours. The paper describes the proposed approach by applying it to an early case study.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... As defined in [1], an Intelligent Tutoring System (ITS) [2], [3] is a software system providing adaptive educational experiences. The main features of an ITS [4] include generation and delivery of (i) learning activities that are aligned to learners' current knowledge and skills status in order to foster meaningful learning ( [5] and [6]), (ii) individualized feedback to stimulate next learning activities and avoid frustration ( [7] and [8]), demotivation and disengagement due to unsuccessful performance, and (iii) guidance (typically in the form of hints for learning activities) that help learners (for instance [9], [10] and [11]) during the execution of their learning tasks [12], [13]. ...
... The case study adopts a dataset gathered in the context of an Italian University and Research Ministry co-funded project, namely INF@NZIA DIGI.Tales 3.6 [54]- [56]. In particular, the dataset has been collected by the University of Salerno [1] in the context of further experimentation, whose results are reported in the work [57] that was partially supported by the same aforementioned project. In particular, that work provides more details about the experimentation of an interactive game-based Edu App, namely Bigfoot the pedestrian, based on Augmented Reality. ...
Article
Full-text available
Pedagogical (Tutor or Tutoring) Models are an important element of Intelligent Tutoring Systems (ITS) and they can be described by sets of (tutoring) rules. The implementation of a Tutoring Model includes both the formal representation of the aforementioned rules and a mechanism able to interpret such representation and execute the rules. One of the most suitable approaches to formally represent pedagogical rules is to construct semantic web ontologies that are highly interoperable and can be integrated with other models in an ITS like the subject domain and the student model. However, the main drawback of semantic web-based approaches is that they require a considerable human effort to prepare and build relevant ontologies. This paper proposes a novel approach to maintain the benefits of the semantic web-based approach in representing pedagogical rules for an ITS, while overcoming its main drawback by employing a data mining technique to automatically extract rules from real-world tutoring sessions and represent them by means of Web Ontology Language (OWL).
... This work proposes an approach to build tutoring models and, in particular, to select the next task difficulty to keep the learner in her ZPD during the interaction with the learning environment. Differently from the existing works [5], [6], [7], [8], this paper provides a novel approach considering both human tutors' competence/experience and individual students' ZPDs. ...
... Observations are gathered across multiple sessions involving several learners and tutors. Whilst the same approach has been used in a previous work [8] to learn rules for feedback and hints, the present work focuses on mining of NTS rules. The game is organized in problems (tasks) and each problem can be solved by correctly executing a sequence of steps. ...
... Moreover, it is also difficult to adapt typical approaches for building personalized e-learning systems, such as the already-mentioned cases of ITSs and ontology-supported e-learning approaches, to be used in POJs. ITSs [17,44,50], are dynamic and adaptive systems for personalized instruction based on the students' characteristics and behavior that involve a combination of various fields such as artificial intelligence, cognitive psychology and educational research. However, they tend to be composed of complex architectures that include a domain model, a student model, a teaching model, and interfaces [36]; therefore their modelling is not possible in current POJs because they do not manage the required information to build them (such as problem difficulties, user learning objectives, etc.). ...
... Consequently, we initially verify the difference between the neighborhoods used in the recommendation generation associated with each user x according to the original binary matrix M (neighbors M x ), and the ). To this purpose, we define the overlapping degree between both sets of neighborhoods for a user x (17), as the ratio between the cardinality of its intersection, and the initial size of the neighborhood (the k value, in this case k = 130). This ratio would indicate a low or a high effect of the proposal in Sections 3.1 and 3.2, in the formation of different and possibly more precise neighborhoods. ...
Article
Full-text available
The use of programming online judges (POJ) to support students acquiring programming skills is common nowadays because this type of software contains a large collection of programming exercises to be solved by students. A POJ not only provides exercises but also automates the code compilation and its evaluation process. A common problem that students face when using POJ is information overload, as choosing the right problem to solve can be quite frustrating due to the large number of problems offered. The integration of current POJs into e-learning systems such as Intelligent Tutoring Systems (ITSs) is hard because of the lack of necessary information in ITSs. Hence, the aim of this paper is to support students with the information overload problem by using a collaborative filtering recommendation approach that filters out programming problems suitable for students’ programming skills. It uses an enriched user-problem matrix that implies a better student role representation, facilitating the computation of closer neighborhoods and hence a more accurate recommendation. Additionally a novel data preprocessing step that manages anomalous users’ behaviors that could affect the recommendation generation is also integrated in the recommendation process. A case study is carried out on a POJ real dataset showing that the proposal outperforms other previous approaches.
... 1. Knowledge tracing systems: The systems in the first category model the mastery level of learners and make predictions about it. Some examples are Bayesian Networks to implement a control shared between the students and the machine to track the process of studying linear equations [6], the use of Artificial Neural Networks in children games to determine the right amount of difficulty for each user [7] and Formal Concept Analysis to determine the type of feedback corresponding to each student when solving a given task [8]. ...
Article
Full-text available
In this paper we propose a theoretical model of an ITS (Intelligent Tutoring Systems) capable of improving and updating computer-aided navigation based on Bloom’s taxonomy. For this we use the Bayesian Knowledge Tracing algorithm, performing an adaptive control of the navigation among different levels of cognition in online courses. These levels are defined by a taxonomy of educational objectives with a hierarchical order in terms of the control that some processes have over others, called Marzano’s Taxonomy, that takes into account the metacognitive system, responsible for the creation of goals as well as strategies to fulfill them. The main improvements of this proposal are: 1) An adaptive transition between individual assessment questions determined by levels of cognition. 2) A student model based on the initial response of a group of learners which is then adjusted to the ability of each learner. 3) The promotion of metacognitive skills such as goal setting and self-monitoring through the estimation of attempts required to pass the levels. One level of Marzano's taxonomy was left in the hands of the human teacher, clarifying that a differentiation must be made between the tasks in which an ITS can be an important aid and in which it would be more difficult.
... For these reasons, we expand the previous definition of the learner's situation by considering the concepts of student motivation and engagement as a fundamental part of the situation. In such a way, a situation-aware adaptive learning system will be able to monitor motivation and engagement and will be able to anticipate a possible learner dropout and react accordingly [8] . ...
Conference Paper
E-learning is becoming increasingly popular among learners of any age, thanks to its many advantages related to the possibility to learn anywhere and anytime and the low costs. Despite this success and diffusion, the student dropout rate of e-learning systems is very high. The main reason is the lack of motivation and engagement of the students with the online course. An adequate design of the e-learning system and its capability to adapt to the characteristics of the learners can reduce the dropout rate. In this work, we propose an approach, based on Fuzzy Cognitive Map, to identify the situation of the learner (mainly in terms of motivation and engagement) and to provide the learners with a set of feedback aiming at improving the retention of the learners in a situation-aware adaptive learning system. A prototypical system is proposed to preliminary verify the feasibility and utility of the approach.
... The choice of using a semantic model to represent the critical information of the system is for supporting the interoperability between the different systems and tools integrated into the learning system, providing them with greater flexibility in the data management, thanks to a unique, shared, formal data model. Moreover, the semantic model is useful to support novel learning analytics techniques (e.g., Formal Concept Analysis [17]) that need a formal representation of the data. Such a formal representation sustains reasoning and inference which can support the decisionmaking processes of teachers and analysts. ...
Conference Paper
The lack of motivation and engagement is recognized as one of the main causes of learners dropping out of e-learning systems. In this paper, an Adaptive Learning System, based on the principles of situation awareness, is proposed to tackle such an issue. The work proposes a situation model based on motivation and engagement. A technique based on Fuzzy Cognitive Map (FCM) has been defined to identify the current situation by tracking the behavior and the interactions of the learner with the system. The FCM drives the process of feedback generation to improve the situation awareness of the learner, and therefore their motivation and engagement. The system has been evaluated using the Situation Awareness Global Assessment Technique, involving students and teachers. The experimental results demonstrate that the system is able to significantly improve the situation awareness of both learners and teachers, reducing the risk of learner dropout.
... Knowledge, goals, interests, experiences and context are some of the information represented in user models. Intelligent tutoring systems handle user profile looking for recognizing student difficulties to offer guidance that facilitates their learning process [9]. ...
Chapter
Full-text available
In communities of practice (CoP), learning occurs through constant interactions of their participants. The social aspect is funda- mental for the construction of knowledge. This work uses semantic web technologies and ontologies to structure and represent the interactions of CoPs participants around a dynamic user profile. This user profile de- scribes a set of dispersed properties and relationships in CoPs, allowing collaborative trajectories recovery in these learning environments.
Article
Full-text available
The diffusion of information and communication technologies, in the last decades, appears to be a great opportunity in teaching and learning processes, not attributable to a simple differentiation of learning supports. Augmented reality (AR), in particular, appears capable of radically transforming learning methods, supporting students to have an educational experience able to involving all the senses, facilitating attention and the storage of content. This paper provides a brief overview on the use of augmented reality in learning, in order to present a didactic re-mediation model and its trial application developed in the context of the Inf@nzia Digitales 3.6 project. The illustrated learning experiences are designed for children from three to six years for the discovery of geometric shapes in a smart city, and their understanding as road signs.
Chapter
In communities of practice (CoP), learning occurs through constant interactions of their participants. The social aspect is fundamental for the construction of knowledge. This work uses semantic web technologies and ontologies to structure and represent the interactions of CoPs participants around a dynamic user profile. This user profile describes a set of dispersed properties and relationships in CoPs, allowing collaborative trajectories recovery in these learning environments.
Technical Report
Full-text available
The tutorial described in this report was conducted at the Interservice/Industry Training Simulation and Education Conference (IITSEC) in Orlando, FL, in December 2014. The purpose of this tutorial is 5-fold: 1) understand the differences between adaptive and adaptable systems; 2) understand the key components of Intelligent Tutoring Systems (ITSs); 3) understand the potential of ITSs as one-to-one tutors and where ITS technologies are most applicable in the training and educational domain; 4) understand the concept of self-regulated learning (SRL); and 5) understand how ITS design can support SRL.
Conference Paper
Full-text available
This work presents an approach to assist teachers, tutors and students from online learning environments. It is a four-steps pro-cess called Pedagogical Recommendation Process that uses the coordi-nated efforts of human actors (pedagogical and technological specialists) and artificial actors (computational artifacts). The process' objective is to find relevant information in educational data to help creating per-sonalized recommendations. Using the process it was possible to detect issues within a learning environment (UFAL Línguas), and discovered why some students were facing difficulties, and what other students were doing in order to succeed in the course. This information was used to personalize pedagogical recommendations.
Conference Paper
Full-text available
Intelligent tutoring Systems (ITSs) adapt content and activities with the goals of being both effective and efficient instructional environments. They have goals for students to be challenged and guided in an optimal way--without being too overwhelmed with difficult material or too bored with easy or repetitive material. We propose a particular definition of the zone of proximal development (ZPD) as a general way to describe what all ITSs try to do, and we propose a foundational analysis of instructional adaptivity, student modeling, and system evaluation in terms of the ZPD. We give an operational definition of the ZPD and give an example of its use, and summarize how instructional methods such as scaffolding can be used to maintain ZPD-learning. We also explain how our definition of the ZPD can lead to a more complete model for efficient and effective instruction than common mastery learning criteria.
Conference Paper
Full-text available
Pedagogical tutorial tactics are policies for a tutor to decide the next action when there are multiple actions available. When the contents were controlled so as to be the same, little evidence has shown that tutorial decisions would impact students’ learning. In this paper, we applied Reinforcement Learning (RL) to induce two sets of tutorial tactics from pre-existing human interaction data. The NormGain set was derived with the goal of enhancing tutorial decisions that contribute to learning while the InvNormGain set was derived with the goal of enhancing those decisions that contribute less or even nothing to learning. The two sets were then compared with human students. Our results showed that when the contents were controlled so as to be the same, different pedagogical tutorial tactics would make a difference in learning and more specifically, the NormGain students outperformed their peers.
Conference Paper
Full-text available
The need for effective tutoring and training is mounting, especially in industry and engineering fields, which demand the learning of complex tasks and knowledge. Intelligent tutoring systems are being employed for this purpose, thus creating a need for cost-effective means of developing tutoring systems. We discuss a novel approach to developing an Intelligent Tutoring System shell that can generate tutoring systems for a wide range of domains. Our focus is to develop an ITS shell framework for the class of Generic Task expert systems. We describe the development of an ITS for an existing expert system, which serves as an evaluation test-bed for our approach.
Article
Full-text available
Recently concept lattices became widely used tools for intelligent data analysis. In this paper, several algorithms that generate the set of all formal concepts and diagram graphs of concept lattices are considered. Some modifications of well-known algorithms are proposed. Algorithmic complexity of the algorithms is studied both theoretically (in the worst case) and experimentally. Conditions of preferable use of some algorithms are given in terms of density/sparseness of underlying formal contexts. Principles of comparing practical performance of algorithms are discussed.
Article
Full-text available
On large problems, reinforcement learning systems must use parameterized function approximators such as neural networks in order to generalize between similar situations and actions. In these cases there are no strong theoretical results on the accuracy of convergence, and computational results have been mixed. In particular, Boyan and Moore reported at last year's meeting a series of negative results in attempting to apply dynamic programming together with function approximation to simple control problems with continuous state spaces. In this paper, we present positive results for all the control tasks they attempted, and for one that is significantly larger. The most important differences are that we used sparse-coarse-coded function approximators (CMACs) whereas they used mostly global function approximators, and that we learned online whereas they learned offline. Boyan and Moore and others have suggested that the problems they encountered could be solved by using actual outcomes (...
Article
Developing intelligent tutoring systems from student solution data is a promising approach to facilitating more widespread application of tutors. In principle, tutor feedback can be generated by matching student solution attempts to stored intermediate solution states, and next-step hints can be generated by finding a path from a student's current state to a correct solution state. However, exact matching of states and paths does not work for many domains, like programming, where the number of solution states and paths is too large to cover with data. It has previously been demonstrated that the state space can be substantially reduced using canonicalizing operations that abstract states. In this paper, we show how solution paths can be constructed from these abstract states that go beyond the paths directly observed in the data. We describe a domain-independent algorithm that can automate hint generation through use of these paths. Through path construction, less data is needed for more complete hint generation. We provide examples of hints generated by this algorithm in the domain of programming.
Chapter
A complex concept lattice can possibly be split up into simpler parts. Here the mathematical model must prove its worth by providing efficacious and versatile methods for the decomposition. Every such decomposition principle can be reversed to make a construction method. Therefore, some of the following subjects will be taken up again in the next chapter with this second focus.
Chapter
Maps between concept lattices that can be used for structure comparison are above all the complete homomorphisms. In Section 3.2 we have worked out the connection between compatible subcontexts and complete congruences, i.e., the kernels of complete homomorphisms. A further approach consists in coupling the lattice homomorphisms with context homomorphisms. In this connection, it seems reasonable to use pairs of maps, i.e., to map the objects and the attributes separately. Those pairs can be treated like maps. We do so without further ado and write, for instance,$$(\alpha ,\beta ):(G,M,I) \to (H,N,J),$$if we mean a pair of maps \( \alpha :G \to H,\beta :M \to N, \) using the usual notations for maps by analogy. This does not present any problems, since in the case that \( G \cap M = + H \cap N \) we can replace such a pair of maps (α,β) by the map $$\alpha \cup \beta :G\dot \cup M \to H\dot \cup N
Article
Tutoring systems are described as having two loops. The outer loop executes once for each task, where a task usually consists of solving a complex, multi-step problem. The inner loop executes once for each step taken by the student in the solution of a task. The inner loop can give feedback and hints on each step. The inner loop can also assess the student's evolving competence and update a student model, which is used by the outer loop to select a next task that is appropriate for the student. For those who know little about tutoring systems, this description is meant as a demystifying introduction. For tutoring system experts, this description illustrates that although tutoring systems differ widely in their task domains, user interfaces, software structures, knowledge bases, etc., their behaviors are in fact quite similar.
Conference Paper
Nowadays, in enterprise environments there is a wide and consolidated utilization of software for the human resource management providing functionalities like organizational management, personnel development, training event management, etc. that lay upon a competencies repository mostly populated through expensive and inefficient data entry activities. The new trends in Web 2.0 see a paradigm namely Enterprise 2.0, for supporting business activities within organizations. Web 2.0 is mainly exploited to sustain collaboration, information exchange and knowledge sharing. This work introduces an agent-based framework for the dynamic refinement of employees' competencies profiles by analyzing and monitoring collaborative activities executed through Enterprise 2.0 tools (e.g. corporate blogs, enterprise wikis, etc.). A fuzzy extension of Formal Concept Analysis model supports the elicitation of implicit knowledge and the content structuring into a conceptual representation. The resulting concept-based organization of initial user-generated content will be exploited to provide automatic hints to human resources (HR) managers in order to support them in making safer decisions that involve employees' competencies.
Conference Paper
Auch erschienen in: Baader, Franz u.a. (Hrsg.): Advances in artificial intelligence. (Lecture notes in computer science ; 2174). Berlin u.a. : Springer, 2001. S. 335-350. ISBN 3-540-42612-4 (The original publication is available at www.springerlink.com) Association rules are used to investigate large databases. The analyst is usually confronted with large lists of such rules and has to find the most relevant ones for his purpose. Based on results about knowledge representation within the theoretical framework of Formal Concept Analysis, we present relatively small bases for association rules from which all rules can be deduced. We also provide algorithms for their calculation. Extern
System of data analysis concept explorer
  • S A Yevtushenko