Agent Technology In E-Learning: Characteristics, Requirements, And Application

Agent Technology Characteristics

E-learning represents one of the most drastic changes in the current era. This affects training and education in the history of education. This adaptive e-learning method helps agents to employ learning material to match their type of personality (Garcia-Cabot, de-Marcos and Garcia-Lopez 2015).

The following report discusses agent technology properties, requirements of e-learning, applying agents to e-learning. Lastly, case studies related to JAM agent case study and BDI model.

2. Agent technology characteristics:

Characteristic of Intelligent Agent:

Every agent is autonomous. This indicates that the agents have controls over their actions. Further, all kinds of artificial agents are driven by specific aims. They have a particular purpose and have been acting according to that meaning. 

 Also, there have been various characteristics of agents through which the agents attain their goals. They are highlighted below.

• All agents are driven but scripts pre-defining actions that must then define those goals of the agents.
• They must have programs and the as long as the program is driven by aims they also have other characteristics of agents.
• Rules should drive agents and they must define goals of agents. Moreover, there has been different goals that are embedded in planning methodologies, and in many situations, they can alter their goals in due time (Arif et al. 2015).

Besides, the agents can sense changes taking place within the environment and responding to those changes. The properties of the agents have been lying at the central part of automation and delegation. Here, for instance, as the assistant is asked that when x takes place, y must be done. Thus the agent always waits for x to take place. The agents continue to work as the user is also gone, that indicates that the agents can run over a server. However, in some cases, the agents run on the user systems.

At multi-agent systems, the agents are social this indicates that they have been communicating with other types of agents. Few artificial agents have been changing and learning the behaviour from prior experiences. Some agents include mobile that indicates that they move from machine to machine to get closer to data that might require for processing and be doing that instead of any delay of a network.

At last some agents have been attempting to be believable so that they get represented as the entity visible or audible to users and might have aspects of personality and emotion (Lavrov et al. 2017).

 The designs and productions have been considering various factors. The characteristics of those agents are identified by considering factors and providing insights into the how intelligent has been affecting society and business overall. The artificial intelligent has comprised various characteristics namely socialability, adaptivity, autonomy and situations.

3. E-Learning requirements:

The various components of e-learning are helpful to consider the multiple elements of presenting, distributing and producing e-learning. Conventional learning needs teachers, classrooms and facilities. This has also included a syllabus, lesson plans, materials and books supporting a course on every subject. The requirements of e-learning are categorized as follows.

• Branding and certification of students and programs
• Learning management
• Hosting technologies and supporting
• Ownership of courseware
• Provisioning of courseware
• Development of content to sources
• Accessing original materials

Here the above elements of creative processes of e-learning have been helpful in analyzing business models and forms. Various institutions and companies have been possessing competencies in some of those sectors. Further, this has been providing rationale from weaknesses and strengths regarding e-learning to develop partnerships for delivering e-learning to clients (Al-Omari, Carter and Chiclana 2015). 

4. Application of Agents to E-Learning:

The demands of e-learning can be understood through its various advantages. The agent to learning has been helping students to work at their speed and skip materials that are known already. Besides, BYOD has been enabling learning at any place and providing platforms for better retention and engagements. It has been eradicating every-day pieces of training in favor of practice accomplished during downtime (Anderson 2016). Moreover, it decreases expenses, travel times and time away from tasks. It has enabled continuous education and mentoring through encouraging through breakdowns with social learning. Further advantages include.

• Realizing cost savings through elimination of dedicated classrooms, computers and instructors.
• Permitting enterprises to create a library of courses serving as the repository of current and new employees.
• Supporting self-starters to pursue career goals.

E-Learning Requirements

However, agents for e-learning have not been devoid of any drawbacks. It has been eliminating some learners ended stricture of classroom instructions. Here, low-tech learners have been facing hard times to navigate pieces of training. The e-learning has needed more upgraded data that is available in workplaces. The experience workers have not been able to perform mobile learning. While introducing e-learning, training departments should be selling the idea to various reluctant employees (Truong 2016).

5. A case study using BDI and JAM agent:

5.1. The BDI agent:

There have been various approaches to suggest different kinds of mental attitudes and relationships. Here the most adopted among them has been the BDI model or “Belief-Desire-Intention”. It is a physical theory related to practical reasoning and providing explanations regarding different intentions, desires and beliefs. Further, a critical assumption of this agent is derived from reasonable rationale comprised of a couple of states (Grajewski et al. 2015). At the initial step, the deliberation of goals and set of desires is achieved as the present situation of the belief of agents. In the second step determination of how those concrete goals are produced due to prior steps can be gained through means of available options for the agents are done.

The three mental attitudes that have been the part of BDI model are analyzed hereafter.

It has been adopted by various software agents and presented as an abstract BDI interpreter and formal theory. This has also comprised of abstract BDI interpreter. This has been the base of every BDI systems whether historical or used at present. Here the interpreter has been operating over different plans, beliefs and goals representing the idea of mentalistic notions having little changes (Klašnja-Mili?evi?, Ivanovi? and Nanopoulos  2015). Here the most significant modification is regarding the goals that have been set of constant concrete desires achieved together. This has been avoiding the necessity of difficult phase to select and execute plans for particular events and goals. In initial system has been implemented with success from interpreter namely PRS Procedural Reasoning Systems. This further got succeeded by the method called dMARs.

There have been a various set of present beliefs that has been denoting the information possesses by agents regarding existing scenarios. Next, it has comprised of belief revision functioning taking a perceptual input and current opinion of agents. Thus in that basis, the new set of beliefs are determined. Next, there has been the function of option generation determining options available to agents.

This has been from present beliefs regarding environments and intentions. These set of current possibilities denoting possible courses of actions are available to agents (Stevenson, Hartmeyer and Bentsen 2017). Again the filter function has been representing the deliberation process and determining plans of agents regarding plans. This has been denoting the present focus stating affairs committed to trying for bringing about. Lastly, that has comprised of action selection functions determining tasks to perform at the base of current intentions.

5.2. The JAM agent:

Application of Agents to E-Learning

The excellent plug-and-play architecture of JAM has been permitting snap-in learning agents. This meta-learning and learning agent has been designed as objects. This has been providing the parent agent class, and all instance agents were implementing the favorite learning algorithms. This is defined as the subclass of that parent class. Among various other definitions inherited by every agent subclasses, the parent agent classes have been providing minimal and simple interfaces that are needed to be compiled by every subclass (Dascalu et al. 2015). As far as learning and meta-learning agents have been conforming to the interface, it gets introduced and utilized immediately within the JAM. More specifically, the JAM has required comprising of various methods to get implemented. The constructor method has possessed no arguments. This has been instantiating agents, providing that are known by the name.

Further, in most of the cases, the initialize method has been inherited by agent subclass from the parent agent class. With the help of this method, JAM has been supplying different needed arguments to those agents (Hameed, Akhtar and Missen 2016). These arguments have been including test and training datasets, a name of dictionary file and filename of output classifier. Here the methods used by JAM have been obtaining the newly built classifiers encapsulated and snapped-in at any data site that has been participating. The remote agent dispatch has been accomplished easily (Allen 2016). 

Thus it is seen that the JAM architecture has been a distributed, scalable, portable and extensible agent-based database system supporting launching meta-learning and learning agents towards distributed sites. This is created on current agent infrastructures that are available over the current Internet (Romero et al. 2015). Moreover, JAM has been able to integrate distributed knowledge and to boost the overall predictive accuracy of numbers of independently learned classifiers around the meta-learning agents. In collaboration with FSTC, those database sites are populated with records of different transactions of credit cards that are supplied by various banks (Poitras and Lajoie 2014). This has been an attempt to prevent and detect fraudulent activities through assimilating learned behaviors and patterns from independent origins.

6. Conclusion:

It can be concluded by saying that lack of personalized learning has been the main shortcoming of conventional e-learning system. To understand the agent technology in e-learning, the report has considered techniques used for developing e-learning processes. It has discussed the two case studies, BDI model and JAM agents improving the learner’s initiative participation. This is helpful with personalized knowledge services. The complexity of analyzing e-learning has been a significant problem. This is needed to be addressed by various educational developers. This possible solution to the part of the problem has been lying in use of software agents for extracting data. This can be done through the proper implementation of the above tools and organize that information intelligently. 

7. References:

Allen, M.W., 2016. Michael Allen’s guide to e-learning: Building interactive, fun, and effective learning programs for any company. John Wiley & Sons.

Al-Omari, M.M., Carter, J. and Chiclana, F., 2015. A Proposed Framework to Support Adaptivity in E-learning Systems.

Anderson, T., 2016. Theories for learning with emerging technologies. Emergence and innovation in digital learning: Foundations and applications, pp.35-50.

Arif, M., Illahi, M., Karim, A., Shamshirband, S., Alam, K.A., Farid, S., Iqbal, S., Buang, Z. and Balas, V.E., 2015. An architecture of agent-based multi-layer interactive e-learning and e-testing platform. Quality & Quantity, 49(6), pp.2435-2458.

Dascalu, M.I., Bodea, C.N., Moldoveanu, A., Mohora, A., Lytras, M. and de Pablos, P.O., 2015. A recommender agent based on learning styles for better virtual collaborative learning experiences. Computers in Human Behavior, 45, pp.243-253.

Garcia-Cabot, A., de-Marcos, L. and Garcia-Lopez, E., 2015. An empirical study on m-learning adaptation: Learning performance and learning contexts. Computers & Education, 82, pp.450-459.

GeeksforGeeks. (2018). Agents in Artificial Intelligence – GeeksforGeeks. 

Grajewski, D., Diakun, J., Wichniarek, R., Dostatni, E., Bu?, P., Górski, F. and Karwasz, A., 2015. Improving the skills and knowledge of future designers in the field of ecodesign using virtual reality technologies. Procedia Computer Science, 75, pp.348-358.

Hameed, M., Akhtar, N. and Missen, M.S., 2016. Role Based Multi-Agent System for E-Learning (MASeL). International journal of advanced computer science and applications, 7(3), pp.194-200.

Ieeexplore.ieee.org. (2018). AI in the IRS – IEEE Conference Publication.

Ieeexplore.ieee.org. (2018). Cyber vulnerabilities on agent-based smart grid protection system – IEEE Conference Publication.

Inf.ufrgs.br. (2018). BDI Architecture | BDI4JADE.

Klašnja-Mili?evi?, A., Ivanovi?, M. and Nanopoulos, A., 2015. Recommender systems in e-learning environments: a survey of the state-of-the-art and possible extensions. Artificial Intelligence Review, 44(4), pp.571-604.

Lavrov, E., Pasko, N., Tolbatov, A. and Barchenko, N., 2017, July. Development of adaptation technologies to man-operator in distributed E-learning systems. In Advanced Information and Communication Technologies (AICT), 2017 2nd International Conference on (pp. 88-91). IEEE.

Márquez, E., Savage, J., Berumen, J., Lemaitre, C., Laureano-Cruces, A.L., Espinosa, A., Leder, R. and Weitzenfeld, A., 2015. A Decision Support System Based on Multi-Agent Technology for Gene Expression Analysis. International Journal of Intelligence Science, 5(03), p.158.

Modernworkplacelearning.com. (2018). 5 Factors driving Modern Workplace Learning – Modern Workplace Learning Magazine.

Poitras, E.G. and Lajoie, S.P., 2014. Developing an agent-based adaptive system for scaffolding self-regulated inquiry learning in history education. Educational Technology Research and Development, 62(3), pp.335-366.

Romero, L., North, M., Gutiérrez, M. and Caliusco, L., 2015. Pedagogically-Driven Ontology Network for Conceptualizing the e-Learning Assessment Domain. Journal of Educational Technology & Society, 18(4).

Stevenson, M.P., Hartmeyer, R. and Bentsen, P., 2017. Systematically reviewing the potential of concept mapping technologies to promote self-regulated learning in primary and secondary science education. Educational Research Review, 21, pp.1-16.

Truong, H.M., 2016. Integrating learning styles and adaptive e-learning system: Current developments, problems and opportunities. Computers in human behavior, 55, pp.1185-1193.

Order an Essay Now & Get These Features For Free:

Turnitin Report

Formatting

Title Page

Citation

Outline

Place an Order