In P. Kommers, A. Dovgiallo, V. Petrushin, & P. Brusilovsky (Eds.), New media and telematic technologies for education in Eastern European countries, Twente University Press, Enschede, 1997, p. 129-140.
Peter Brusilovsky
School of Computer Science, Carnegie Mellon University
Pittsburgh, PA 15213, USA
E-mail: plb@cs.cmu.edu
Abstract: In this paper we discuss the problems of building educational and authoring systems which integrate the capability of intelligent tutoring systems, hypermedia systems, and learning environments. We describe our experience and lessons learned on development of a number of integrated systems. We also present our approach for building authoring systems for such integrated courseware and provide an extended analysis of related works giving some fruitful ideas for this area of research.
Three classes of advanced educational computer systems: intelligent tutoring systems (ITS), hypermedia systems, and learning environments. are no longer considered as contradictory ways of using computers in education. Recent research has demonstrated that hypermedia and learning environments can provide the basis for an exploratory learning system but that by itself such a system is insufficient, needing to be supplemented by more directed guidance (Hammond, 1989; Shute & Glaser, 1990). The guidance is often an important ingredient of effective learning from hypermedia and exploratory environments. The guidance can be provided by an intelligent tutoring component. Only a system which integrates the capabilities of an intelligent tutor, a learning environment and a hypermedia system can support the full continuum of learning styles from unstructured instructional environment to guided curriculum for learning various kinds of knowledge (Tab.1)
Table 1. Learning styles and knowledge supported by some known kinds of systems
Knowledge vs. Control |
Procedural knowledge |
Declarative knowledge |
System-controlled (Tutoring) |
ITS ("Coach") |
ITS ("Tutor") |
Student-controlled (Learning) |
Learning environment |
Hypermedia |
For a number of years our group based in the Moscow State University and the International Center for Scientific and Technical Information (ICSTI), have been investigating the problem of creating integrated intelligent learning environments. We have designed and tested the Intelligent Tutor, Environment and Manual for Introductory Programming (Brusilovsky, 1992b) which demonstrates an approach to integrating an ITS and a learning environment for programming. Later we have tried to apply the same approach to some other domains. We have developed ILE for mathematics (Brusilovsky V., 1993), print formatting language (Pesin & Brusilovsky, 1992), and geography (Brusilovsky & Gorskaya-Belova, 1992). The first versions of these ILE contained an intelligent tutor and a learning environment integrated together on the basis of the domain and student models, but did not contain hypermedia components. More recently we have tried to extend our approach by integrating the hypermedia technology into our existing ILE. We have build hypermedia-based versions of our ILE for print formatting language (Brusilovsky & Pesin, 1994) and geography (Brusilovsky & Zyryanov, 1993). This experience led us to form an extended approach and a framework to building ILE based on the integration of hypermedia and intelligent tutoring technologies. The framework is used currently by our group as a basis for design and re-design of several ILE and authoring tools.
In this paper we are trying to analyze and generalize our experience in the development of hypermedia-based ILE. First we describe the architecture of the original ITEM/IP system placing emphasize on the ways used to support a combination of system-controlled and student-controlled modes. Then we suggest general ideas on integrating hypermedia technology into the original approach and describe briefly our recent hypermedia-based ILE which are based on these ideas. After that we summarize our experience and the main features of our updated approach to building integrated ILE based on hypermedia and intelligent tutoring technologies and discuss its application for authoring. A special section is devoted to an extensive analysis of related works, which presents the background of our research and gives some more fruitful ideas for building integrated authoring tools. We conclude with a summary of the potential benefits provided by the integration of hypermedia and intelligent authoring technologies.
ITEM/IP stands for Intelligent Tutor, Environment and Manual for Introductory Programming (Brusilovsky, 1992b). ITEM/IP is an ILE for introductory programming based on the mini-language approach (Brusilovsky et al., 1994). It was designed in 1985 to support first year students of the Moscow State University. The student can use ITEM/IP in a system-controlled mode or in two different learner-controlled modes:
1. Novice programming environment. To work with the mini-language programs the student can make use of a complete set of tools for program design and debugging. One of the functions of the environment is to display visualizations of the student's programs. With the help of the environment the student can observe various programs "at work", experiment with them and gradually learn from experience, observations and mistakes.
2. Intelligent tutoring. At any moment of work the student can get the next "best" teaching operation from the teaching component. The teaching component applies an imbedded tutoring strategy and use the domain knowledge and the student model to offer the student the best teaching operation. Five kinds of operations are used: presentation of new concept or construct, construct example, test, problem solution example, or programming problem to solve.
3. On-line manual for student-driven learning and repetition. At any moment of work the student can use the "learn" and "repeat" hierarchical menus to access all ready-to-be-applied teaching operations (which have all the prerequisites learned), as well as all previously learned material: presentation of any known concept, demonstration of all learned examples, and analysis of any explained or solved problem. This mode offers a reference access to all learned and ready-to-be-learned material of the course, thus supporting student-controlled learning, on-line help and example-based programming.
The three main components of ITEM/IP are the novice programming environment mentioned above, the tutoring component which supports both intelligent tutoring and on-line manual mode, and the knowledge base which serves as a kernel of the system. The knowledge base includes the domain model, the student model, and the knowledge base of learning material. The domain model in ITEM/IP is a network with nodes corresponding to programming concepts and language constructs and with links reflecting several kinds of relations between nodes ("is-a", "part-of" and prerequisite). The overlay student model reflects by a set of integer counters the extent to which the student has mastered the concept. Student model is always kept up-to-date and supports adaptive work of all modules.
All the learning material is stored as frames in the knowledge base of learning material. There are three kinds of frames: concept frames, example frames and problem frames. The presentation module uses these frames to generate the five kinds of teaching operations mentioned above. Example and problem frames are related to the domain model by a special slot called "spectrum" which is a weighted list of concepts related to the given frame. Using the student model and the spectra the tutoring component can compile at any moment the list of relevant teaching operations (with no unknown concepts in its spectra).
ITEM/IP provides both instructional and reference access to the teaching material. Generally, these two methods of access require different structuring and presentation of the material. That is why many good books on programming languages consist of two independent parts: the tutorial and the reference manual. In ITEM/IP project we tried to make an integrated tutorial and manual by providing adaptive presentation of the same teaching material. Briefly speaking, the presentation of teaching material is adapted to the current level of student knowledge, so a repeated explanation is usually more concise and complete than the explanation presented when a concept is initially introduced. This become clear from the following example.
The textual information which is stored for the given concept can be divided into a sequence of text fragments. Each fragment has a condition which addresses the knowledge level of the given and related concepts. While producing a description of the concept, the presentation module presents only the fragments with true condition. Usually, the more the concept is learned, the more concise descriptions are presented. At the same time the description gives more and more complete information about related concepts. The textual part of the description of is followed by a generated list of related concepts, giving the names of related concepts for each possible kind of relationships. To avoid confusion only the relation with already known nodes are presented, so the more neighbors that are known, the more complete information presented. Thus even in the student-controlled mode the system adapts its work to the student and supports the navigation through the learning material by protecting the student from bumping into not-ready-to-be-learned or unknown material. To protect the student from cognitive overload "learn" menus never list concepts with unlearned prerequisites or concepts from next goal subsets.
ITEM/IP system was tested in the classroom. We were satisfied with the approach of integrating student-controlled and system-controlled modes of learning on the basis of the domain and student models. The suggested approach was later generalized and used in several other ILE. At the same time we were never satisfied with hierarchy menus as the only student-controlled way to access the teaching material. This interface appears to be too complicated for novice programmers. For example, when learning or repeating a description of a concept, the student was provided with the references to all previously learned related concepts. However, to refresh their knowledge the students need to re-enter the repetition mode from the top and access the related information by a three-level menu.
Recently the influence of hypermedia technology led us to redesign our approach and the framework. We decide to try the hypermedia paradigm for supporting student-controlled access to all the learning material. Our goal was to integrate the hypermedia component into the general architecture of our ILE based on the domain and student models. The next section reports our recent experience on integrating a hypermedia component into our intelligent learning environments - ISIS-Tutor and ITEM/PG.
We found that the model-based approach we have used in ITEM/IP and similar ILE is very suitable for the integration of intelligent tutoring and hypermedia technologies. We require only "two steps" to integrate a hypermedia network into our ILE. The first step is to take the domain model network of an ILE as a basis for the hypermedia network. Actually, different ideas about creating the hypermedia network from the domain concept network are quite popular and have been discussed elsewhere (Hendley, Whittington & Jurascheck, 1993). What we suggest is to design the main part of the hypermedia network just as the visualized (and externalized) domain network. Each node of the domain network should be represented by a node of the hyperspace, while the links between domain network nodes constitutes main paths between hyperspace nodes. Moreover, each teaching operation should be also represented as a node of the hyperspace and connected to all domain concepts listed in its spectrum. Thus the structure of the overall hyperspace resembles the pedagogic structure of the domain knowledge.
The second step concerns the content of hypermedia pages. We decide not to duplicate the learning material in the form of static hypermedia pages, but have the hypermedia pages generated from the material stored in the knowledge base of learning material. Thus, each concept and each teaching operation will have a hypermedia page (or block of pages) as an external representation, and have a frame as an internal representation. What the student will see on the screen visiting a hypermedia node is really generated from the corresponding frame by a special program which can take into account the student current state of knowledge. This approach not only saves page design time but also provides space for adaptation. We used the above ideas in our recent works on adaptive hypermedia components for ILE. The following two subsections report how the domain and student models are used in ISIS-Tutor and ITEM/PG systems to generate the 'adaptive pages' for hypermedia nodes.
ITEM/PG (Brusilovsky & Zyryanov, 1993) is an ILE for 13-14 years old students taking a course in the physical geography of oceans and continents. The pedagogic goal of the work with ITEM/PG is to learn the relationships between different components of a natural complex of an island, located somewhere in the north part of the Atlantic ocean. The island is characterized by several components of its natural complex as position, origin, climate etc. Each component has several possible values. For example, the options for the origin of the island are continental, volcanic, roof or atoll. The values of the components are related to each other. For example, the climate depends on the position of the island.
The student can work with ITEM/PG in exploratory mode, browsing the hyperspace or experimenting with the environment. At any given moment the student can apply to the tutor for the next best teaching operation, and then can follow the tutor's suggestion or give up. An interesting feature of ITEM/PG is that the best problem is not selected from the pre-stored set of problems, but generated by the tutor. According to our general approach, the hypermedia network is just the visualized domain network. There are three main kinds of hyper-nodes: component, value and rule. Each kind of hyper-nodes has a special screen representation which is generated from the corresponding frame by a special program. For example, the program for the "value" kind of node generates the window which contains the name of the value, the description, the icon and the hyper-links to the component node and related rule nodes.
A new feature of ITEM/PG is adaptive navigation support. The student model is used by the hypermedia component to adapt the screen representation of hyper-links to related nodes. Applying the student model, the hypermedia component distinguishes four knowledge states for each hyper-node: not-ready-to-be-learned (i.e. the nodes which have unlearned prerequisite nodes), ready-to-be-learned, known and well-known. Thus, at any moment the hyperspace is divided implicitly into several "zones". In particular, the ready to be learned nodes form the "zone of proximal development" (Vygotsky, 1978). Our idea is that different zones have different meanings for the student and marking these zones visually would help the student in hyperspace navigation as bookmarks helps the student with textbooks. To mark the zones the hypermedia component just marks the hyper-links of each node by four different colors. For example, the links to the nodes which are not-ready-to-be-learned are gray so as not to distract the student.
Another important new feature of ITEM/PG is an improved student-model centered architecture of the system. According to this architecture an ILE is an open set of components around the linked domain and student models. In this architecture all the modules can not only use but also update the student model (in classic ILE architecture only the tutor or student modeling component can do it). More extended description of the proposed architecture can be found in (Brusilovsky, 1993). What is important for hypermedia is that not only does the hypermedia component of ITEM/PG use the student model for adaptation, but also any student's 'visit' to a particular node of the hyper-network is reflected by the student model. Thus ITEM/PG provides a good example of how the hypermedia component can be integrated into the structure of an ILE.
ISIS-Tutor (Brusilovsky & Pesin, 1994) is an ILE to support learning the print formatting language of well known information retrieval system CDS/ISIS. This system is being supplied by UNESCO and used widely in ICSTI and other information centers in the world. Print formatting language is a key to many CDS/ISIS operations and mastering the language is important for the effective use of the system.
The architecture of ISIS-Tutor resembles the architecture of ITEM/IP in most details. It contains an environment to experiment with the language and a tutoring component which deals with three kinds of teaching operations: concept presentations, examples and problems. The domain model which is a network of 69 concepts and constructs is twice as complex than the one in ITEM/IP. The overall 'space' of teaching material is bigger as well. The adaptive hypermedia component of ISIS-Tutor was implemented according our general approach to integration of ITS and hypermedia. This component is very similar to the adaptive hypermedia component of ITEM/PG. It can distinguish the same four knowledge states for each hyper-node as ITEM/PG and use colors to annotate the hyper-links according to their current state.
Figure 1. Adaptive annitation of links in ISIS-Tutor.
What is special for ISIS-Tutor is the application of two different techniques for adaptive navigation support. In addition to adaptive annotation this system applies adaptive hiding of non-relevant links. In fact, the earlier version of ISIS-Tutor (Brusilovsky et al., 1993) used hiding as the primary technique for adaptive navigation support. In the current version hiding is implemented on the basis of an annotating technique and can be considered as a special variant of it. It also takes into account the educational and goal status of related nodes. The idea of hiding in ISIS-Tutor is to reduce the cognitive load by hiding from the student all links to the nodes which the student is "not expected to learn". There are two kinds of these links in ISIS-Tutor: links to not-ready-to-be-learned nodes and links to the nodes which are outside the current educational goal. In pure "annotating" mode these links were annotated by some non-distractive color (figure 1). In "hiding" mode these links are hidden: they are removed from any menus and their position in the menus is occupied by the next visible links (figure 2). As a result all menus of links become much shorter for a novice who gets most of the links hidden. For example, the full index of concepts takes two screens in ISIS-Tutor. In the hiding mode the novice starts with the index containing just two links, and during the first sessions the growing index still fits one page (figure 2). Note that "hiding" mode in ISIS-Tutor is more advanced then hiding in other adaptive hypermedia systems (Pérez et al., 1995). It is actually a combination of hiding and annotating, because learned, in-work and ready-to-be-learned nodes are still annotated differently as in normal annotating mode.
Figure 2. Adaptive annotation of links from Index page. Links to not-ready-to-be-learned concepts are hiden.
We decided to design ISIS-Tutor and ITEM/PG hypermedia-based ILE to test our general ideas about integrating ILE and hypermedia and to check some particular features of a hypermedia-based ILE. This work was really creative for us. First of all, we get the feeling that our general "two steps" approach is really good and we are moving in the right direction. The integration looks quite natural and really adds new quality to our old systems. Second, all the new features we have tried in ISIS-Tutor and ITEM/PG (adaptive navigation support, hypermedia access to all teaching operations, student model-centered internal architecture) appears to be very useful.
The most important lesson we learned from the experience with ISIS-Tutor and ITEM/PG is that hypermedia really brings new quality. It means that simple mapping of a pre-hypermedia system structure to the hypermedia-based system structure is usually not the best solution. Some part of the system should be redesigned and some new features can be added which can never be useful without the hypermedia. For example, when designing the first hypermedia-based version of ISIS-Tutor we applied the same method of student-controlled access to learning material as in ITEM/IP: two hierarchical menus "learn" and "repeat" for new and old material respectively. However, it appears that the global menu for "new" concepts, the global menu for "old" concepts, and local lists of related concepts adaptively generated for each hyper-node give three different views to the same domain knowledge. It makes the understanding of the domain knowledge structure difficult for the students. In the current version of ISIS-Tutor we have replaced "old" and "new" menus by a classic hypermedia feature: an index of hyper-nodes. This index is adapted by annotating and hiding techniques to be consistent with the way of presenting a list of related concepts in each hyper-node.
Finally, we should mention that there are some design problems for us which needs additional investigations. For example, we use different methods to deal with links from the given node to not-ready-to-be-learned nodes: these links can be hidden as in ITEM/IP or ISIS-Tutor, or marked by a non-distractive color as in ISIS-Tutor and ITEM/PG. We can't say now which way is better. By hiding these links completely we can reasonably limit the cognitive load of the student. However hiding any links looks unnatural for hypermedia. It can form incorrect mental model of the domain knowledge structure. Such adaptation is also more intrusive than just dimming these links. We are now running some classroom studies to evaluate and compare adaptive annotation and adaptive hiding techniques. The results of the first study (Brusilovsky, Pesin, 1995) have shown that both ways of adaptive navigation support significantly reduced navigation effort, but have not shown any difference between annotation and hiding.
After several years of work with various ILE structured according to our "pre-hypermedia" approach and recent trials to integrate it with hypermedia technology we feel that we have now a consistent and handy approach to building hypermedia-based intelligent learning environments which supports both system-controlled and student-controlled modes of learning. This approach was presented step by step in "genetic" form in the above sections. Here we briefly repeat the main features of the approach.
We will be working further testing and improving this approach. New versions of ITEM/IP and ISIS-Tutor are developing now in complete accordance with the presented approach and we are developing some new systems for different domains. The approach we are using is now more or less stable. Thus when we are designing a new system according to our approach, the serious part of work is just repeating and re-coding the same things as we did designing more old systems. Of course, we have some re-usable parts (as sequencing module) which are used by several systems, but now we are going further. We have started recently a project devoted to the design of an authoring system based on our approach for creating hypermedia-based ILE for World Wide Web (Brusilovsky, 1995). Our approach is a good basis to start, but we don't expect to limit ourselves to its current state. While working on our previous ILE and while thinking about an authoring system we have carefully analyzed the existing literature on creating authoring systems for ITS/ILE and hypermedia paradigms and also on integrating hypermedia and ILE technology. The next section gives a brief analysis of works that can serve as a source of good ideas for design of hypermedia-based ILE and authoring tools for it.
We outline the following local areas of research as sources of good ideas for the integration of hypermedia and intelligent tutoring technology:
Going from ITS authoring, we should mention such authoring system as IDE (Russel, Moran and Jordan, 1988), ECAL (Elsom-Cook & O'Malley, 1990) and the works of Vassileva (1990). These authoring systems are based on both CAI and ITS paradigms. The teaching material is not generated as in 'orthodox' ITS, but stored in frames. However, this representation is based on an explicit domain model network, so all these frames can be selected intelligently. Some amount of pedagogic knowledge is also explicitly represented in these systems either in the form of strategies or pedagogic attributes which can be used by an instructional planner. The above works are well-complimented by works of Murray and Woolf (1992) on KAFITS, where special attention was paid to domain and pedagogic knowledge acquisition and representation, and the works of Major and Reichgelt (1992), Marcenac (1992), and Van Marcke (1992) who pay special attention to teaching strategies representation and use in ITS shells. All the above ideas can be easy integrated with hypermedia technology. In fact, some of the mentioned ITS shells already use hypermedia for some local purposes.
Moving on from CAI and hypermedia authoring side we should mention the works on indexing the teaching material. In our approach all teaching operations have a weighted index (spectrum) of related concepts which is a basis for knowledge-based sequencing of teaching operations. In the hypermedia area an idea of indexing was elaborated by Mayes, Kibby and Watson (1988) in StrathTutor system. They stress additional preference of indexing the frames of learning material - the possibility to indicate related pairs of frames not by tedious manual linking of pieces of learning material together, but dynamically, on the basis of similarity of corresponded sets of topics. Indexing of learning material is used also in authoring systems based on courseware re-use paradigm. This work was pioneered by Osin (1976), who suggest a framework for indexing CAI frames by a set of topics which it covers. Such indexed sets of frames was not related with any pre-scribed order of presentation. They can be accumulated, stored in special libraries and re-used by different authors to create their own courses. In the multimedia field a similar idea of a re-usable database of multimedia learning material indexed by topics and keywords is elaborated by Olimpo et al (1990).
We think that an indexed database of learning material is a good step towards knowledge-based hypermedia authoring system. An example of such a system is HITS (Tang, Barden & Clifton, 1990). As any authoring system based on re-use paradigm, HITS separates authoring of teaching material (creating hyperbooks) and course design. The duty of the course designer is to design a syllabus (a hierarchy of subjects with topics attached to it), to choose the hyperbooks to be used and to specify the goal level of student knowledge (in terms of the student model). HITS will manage the course itself, selecting or suggesting the relevant information and tests from the chosen hyperbooks according the syllabus. A more advanced framework for using a database of learning material by an ITS shell is suggested in (Diessel, Lehmann & Vassileva, 1993).
Another important source of ideas are various works on integrating hypermedia and artificial intelligence technology. One of the most interesting examples of real integration of AI and hypermedia technologies is to extend hypermedia navigation by providing intelligent pseudo-links (Beltran & Peninou, 1992; Briggs, Tompsett & Oates, 1991; Nicolson & Tomlinson, 1991, Kushniruk & Wang, 1994).
The fourth source of ideas is the works on adaptive hypermedia which is a recognized area in the domain of adaptive interfaces. Adaptive hypermedia combines free browsing with individual support and can support the full continuum of learning modes between pure system-controlled and student-controlled. Here we just mention two direction of works on adaptive hypermedia: the works on adapting the content of a hypermedia page (Böcker, Hohl & Schwab, 1990; Beaumont, 1994) and the works on adaptive navigation support (de La Passardiere & Dufresne, 1992; Kaplan, Fenwick and Chen, 1993; Pérez et al., 1995). We consider the ideas of adaptive hypermedia as very important and use both ways of adaptation in our approach. A review of works on adaptive hypermedia can be found in (Brusilovsky, 1994; Beaumont & Brusilovsky, 1995)
In this paper we present step by step our approach to building a hypermedia-based intelligent learning environment, discuss its application for the authoring of courseware and provide a brief analysis of related ideas which can be used in further works on authoring tools for "intelligent hypermedia" courseware. In conclusion we want to summarize some preferences for integrating hypermedia and intelligent tutoring technologies.
Intelligent tutoring technology adds support for those learners who can not manage their own learning, in particular those who "get lost". The hyperspace of learning material based on domain knowledge analysis gets well structured. A number of links can be created dynamically decreasing authoring time. The application of explicit domain and student models opens the doors for adaptive hypermedia which is more supportive for the learner. On the other side, hypermedia technology extends ITS and ILE with powerful and natural way of student-controlled learning. It extends the space for exploration for good students and solves the problem of inadequate choice of teaching operation. Additional reasons for the integration of these technologies are making authoring less tedious, involvement of a human teacher, and supporting of known advanced hypermedia technologies as developing hypermedia on the base of network domain models, or as indexing and re-use of learning material. We think that integrating hypermedia and intelligent tutoring technologies is a productive approach to developing more powerful learning systems and authoring tools.
I would like to thank Tom Boyle for his comments on the earlier version of this paper. Part of this work was supported by Alexander von Humboldt-Stiftung Fellowship to the author.
Beaumont I. (1994). User Modeling in the Interactive Anatomy Tutoring System ANATOM-TUTOR. User models and user adapted interaction, 4(1), 21-45.
Beaumont I. & Brusilovsky P. (1995). Educational applications of adaptive hypermedia. In K. Nordby, P. Helmersen, D. J. Gilmore, & S. A. Arnesen (Eds.), Proceedings of Interact'95 (pp. 410-414). London: Chapman & Hall.
Beltran Th. and Peninou A. (1992) An object-oriented approach to produce educational hypermedia software. In Tomek I. (ed.) Computer Assisted Learning. Proceedings of the 4th International Conference, ICCAL'92. Springer-Verlag, Berlin. 111-123.
Böcker H.-D., Hohl H. and Schwab T. (1990) deg.pADAPTer - Individualizing Hypertext. In Diaper D. et al (ed.) INTERACT'90. Proceedings of the the IFIP TC13 Third International Conference on Human-Computer Interaction. North-Holland, Amsterdam. 931-936.
Briggs J.H., Tompsett C. and Oates N. (1991) Guiding learners through a medical hypertext database using knowledge based rules. In Bottino R.M., Forcheri P. and Molfino M.T. (eds.) Knowledge Based Environments for Teaching and Learning. Proceedings of the Sixth International PEG Conference. Genova. 571-577.
Brusilovsky P.L. (1992a) A framework for intelligent knowledge sequencing and task sequencing. In Frasson C., Gauthier G. and McCalla G.I. (eds.) Intelligent Tutoring Systems. Proceedings of the Second International Conference, ITS'92. Springer-Verlag, Berlin. 499-506.
Brusilovsky P.L. (1992b) Intelligent Tutor, Environment and Manual for Introductory Programming. Educational and Training Technology International, 29(1), 26-34
Brusilovsky P. (1993) Student as user: Towards an adaptive interface for an intelligent learning environment. In Brna P., Ohlsson S. and Pain H. (eds.) Proceedings of the AI-ED'93, World Conference on Artificial Intelligence in Education. AACE, Charlottesville. 386-393.
Brusilovsky P. (1994). Adaptive hypermedia: the state of the art. In P. Brusilovsky (Ed.), Proceedings of East-West International Conference on Multimedia, Hypermedia and Virtual Reality (pp. 24-29). Moscow: ICSTI.
Brusilovsky, P. (1995) Intelligent tutoring systems for World-Wide Web. in Proceedings of Third International WWW Conference. Darmstadt: Fraunhofer Institute for Computer Graphics. 42-45.
Brusilovsky P., Kouchnirenko A., Miller P. and Tomek I. (1994) Teaching programming to novices: A review of approaches and tools. In T.Ottman, I.Tomek (eds.) Proc.of ED-MEDIA'94 - World conference on educational multimedia and hypermedia. Vancouver, Canada, June 25-30, 1994. 103-110
Brusilovsky P.L. and Gorskaya-Belova T.B. (1992) The Environment for Physical Geography Teaching. Computers and Education, 18(1-3), 85-88.
Brusilovsky P. and Pesin L. (1994) ISIS-Tutor: An adaptive hypertext learning environment. To be publsished in Proc. of Japanese-CIS Workshop on knowledge-based software engineering. Tokyo.
Brusilovsky P. and Pesin L. (1995) Visual annotation of links in adaptive hypermedia. in Conference Companion of CHI'95. Washington, ACM. 222-223.
Brusilovsky P., Pesin L., Zyryanov M. (1993) Towards an adaptive hypermedia component for an intelligent learning environment. In Bass L.J., Gornostaev J. and Unger C. (eds.) Human-Computer Interaction. Lecture Notes in Computer Science 753, Springer-Verlag, Berlin, 348-358
Brusilovsky P. and Zyryanov M. (1993) Intelligent tutor, environment and manual for physical geography. In PEG'93. Proceedings of the Seventh International PEG Conference. Edinburgh. 63-73.
Brusilovsky V. (1993) Task sequencing in an intelligent learning environment for calculus. In PEG'93. Proceedings of the Seventh International PEG Conference. Edinburgh. 57-62.
Diessel T., Lehmann A. and Vassileva J. (1993) Individualised course generation using a "Slightly intelligent" CAL system. In PEG'93. Proceedings of the Seventh International PEG Conference. Edinburgh. 63-73.
de La Passardiere B. and Dufresne A. (1992) Adaptive navigational tools for educational hypermedia. In Tomek I. (ed.) Computer Assisted Learning. Proceedings of the 4th International Conference, ICCAL'92. Springer-Verlag, Berlin. 555-567.
Elsom-Cook M.T. and O'Malley C. (1990) ECAL: Bridging the gap between CAL and intelligent tutoring systems. Computers and Education, 15(1), 69-81
Hammond N. (1989) Hypermedia and learning: Who guides whom?. In Maurer H. (ed.) Computer Assisted Learning. Proceedings of the 2-nd International Conference, ICCAL'89. Springer-Verlag, Berlin. 167-181.
Hendley R.J., Whittington C.D. and Jurascheck N. (1993) Hypermedia generation from domain representations. Computers and Education, 20(1), 127-132.
Kaplan C. and Fenwick J. and Chen J. (1993) Adaptive hypertext navigation based on user goals and context. User models and user adapted interaction, 3(2).
Kushniruk A. & Wang H. (1994). A hypermedia-based educational system with knowledge-based guidance. In T.Ottman & I.Tomek (Eds.), Proceedings of ED-MEDIA'94 - World conference on educational multimedia and hypermedia (pp. 335-340). Charlottesville: AACE.
Major N. and Reichgelt H. (1992) COCA: A shell for intelligent tutoring systems. In Frasson C., Gauthier G. and McCalla G.I. (eds.) Intelligent Tutoring Systems. Proceedings of the Second International Conference, ITS'92. Springer-Verlag, Berlin. 523-530.
Mayes J.T., Kibby M.R, Watson H. (1988) StrathTutor: The development and evaluation of a learning-by-browsing on the Macintosh. Computers and Education, 12(2), 221-229.
Marcenac P. (1992) An authoring system for ITS which is based on a generic level of tutoring strategies. In Tomek I. (ed.) Computer Assisted Learning. Proceedings of the 4th International Conference, ICCAL'92. Springer-Verlag, Berlin. 428-440.
Murray T. and Woolf B.P. (1992) Tools for teacher participation in ITS design. In Frasson C., Gauthier G. and McCalla G.I. (eds.) Intelligent Tutoring Systems. Proceedings of the Second International Conference, ITS'92. Springer-Verlag, Berlin. 593-600.
Nicolson R.I. and Tomlinson P. USHIR: An intelligent knowledge-based hypermedia system. Hypermedia, 3(1), 1-33
Olimpo G. et al. (1990) On the concept of database of multimedia learning material. In Proceedings of the World Conference on Computers and Education. North Holland, Amsterdam. 431-436.
Osin L. (1976) SMITH: How to produce CAI course without programming. International Journal on the Man-Machine Studies, 8, 207-241
Pesin L.A., Brusilovsky P.L. (1992) Adaptive task sequencing in ISIS-Tutor. In: Abstracts of East-West Conference on Emerging Computer Technologies in Education, Moscow, 6-9 April 1992. - Moscow, 176.
Pérez, T., Lopistéguy, P., Gutiérrez, J., & Usandizaga, I. (1995). HyperTutor: From hypermedia to intelligent adaptive hypermedia. In H. Maurer (Eds.), Proceedings of ED-MEDIA'95 - World conference on educational multimedia and hypermedia (pp. 529-534). AACE, Charlottesville.
Russell D.M., Moran T.P. and Jordan D.S. (1988) The instructional-design environment. In Psotka J., Massey D.L. and Mutter S.A. (eds.) Intelligent tutoring systems: lessons learned. Lawrence Erlbaum Ass, Hillsdale. 203-228.
Shute V.J. and Glaser R. (1990) A large-scale evaluation of an intelligent discovery world: Smithtown. Interactive Learning Environments, 1(1), 51-77
Tang H., Barden R. and Clifton C. (1990) A new learning environment based on hypertext and ITS technology. In Proceedings of the International conference on advanced research on computers in education. Tokyo.
Van Marcke K. (1992) Instructional expertise. In Frasson C., Gauthier G. and McCalla G.I. (eds.) Intelligent Tutoring Systems. Proceedings of the Second International Conference, ITS'92. Springer-Verlag, Berlin. 234-243.
Vassileva J. (1990) An architecture and methodology for creating a domain-independent, plan-based intelligent tutoring system. Educational and Training Technology International, 27(4), 386-397
Vygotsky L.S. (1978) Mind in society. Harvard University Press, Cambridge.