Introduction

The aim of AVANTI, a collaborative R&D project partially funded by the ACTS programme of the European Commission, is the development and evaluation of a distributed information system that provides hypermedia information about a metropolitan area (e.g., public services, transportation, buildings) for a variety of users with different needs (e.g., tourists, citizens, travel agency clerks, elderly people, blind persons, wheelchair-bound people, users with (slight) forms of distrophy).

In AVANTI, the methods and tools developed in the context of adaptive and adaptable systems in the last years, the availability of standardized software components from the area of the World-Wide Web and the widespreadness of computers which are interconnected by metropolitan area networks will be joined in a fruitful way in order to get an ubiquitous information service which is able to take the aims, interests, experiences, and abilities of its users into account.

User Needs

Investigations with respect to user needs have shown a considerable heterogenity for the user groups in concern such as blind persons, wheelchair-bound people, users with (slight) forms of distrophy, travel agency clerks, and tourists; moreover, individual differences in user needs have also been encountered. Some examples might illustrate this:

• For blind users, the modality of information presented must be changed with respect to tactile and/or audio output. Moreover, additional orientation and navigation aids (e.g., table of contents, indices) are helpful for this user group.
• For wheelchair-bound users, additional information concerning accessibility of premises (e.g., existence and dimensions of ramps and/or lifts, kind and width of doors) is valuable and should therefore be provided.
• For users with (slight) forms of distrophy, the man-machine interface (i.e., interaction objects and associated manipulation techniques) should be adapted accordingly.
• For laypersons like a tourist in a booking scenario, a technical term like 'check-in deadline' should be augmented by an explanation. This is (normally) not necessary for domain experts like a travel agency clerk.
• For users interested in paintings, or works of art in general, possibly interesting detail information like the assessment of a painter within a virtual museum should be provided.
• For users with low-bandwidth network access (e.g., via a modem), information that requires high bandwidth like videos and high-resolution pictures should be replaced by less demanding, but nevertheless appropriate, equivalents.

When starting to think about implementation issues it becomes obvious that all these needs can hardly be addressed within the scope of a single project. Consequently, we are currently consolidating and focusing these user requirements, mainly with respect to the field trials that are planned in the third year of the project. The elicited user needs are transformed into a requirements specification for the envisaged information system, thereby taking the dedicated user groups and also resource limitations into account. A complementary task which is currently carried out is to search for similar projects and/or already existing products in the area of the World-Wide Web.

Adaptability and Adaptivity

Tailoring information systems in order to correspond to different user needs can be done manually by the user or automatically by the system. Systems that allow the user to change certain characteristics of the system in order to adapt systems behaviour to his/her preferences are called adaptable, while systems that change their behaviour automatically according to (anticipated) user needs are called adaptive [Opp94].

Both features, adaptability and adaptivity, will be implemented in the AVANTI system. User interface adaptability and adaptivity will address issues like (special) I/O devices (e.g, macro mouse, Braille display, speech synthesizer), interface primitives (e.g., visual and non-visual interface objects), and associated interaction techniques (e.g., visual and non-visual interaction), thereby ensuring the overall access to the information system. Moreover, adaptibility and adaptivity on the content level (i.e., directly referring to hypermedia objects, their presentation and interrelation in hypermedia space) will directly contribute to the usefulness of the information provided and will therefore offer an added-value to the user. For the rest of the paper, we are focusing on content adaptivity because (1) it's not widespread in nowadays systems, and (2) it's organically complementing efforts that aim to offer access to the World-Wide Web for everyone (i.e., including users with disabilities).

User Modeling

In order to provide adaptive system behaviour a so-called 'user model' has to be set up and maintained by the information system. A user model contains assumptions which represent interesting characteristics of an individual user like background knowledge, preferences, interests, and abilities. Different methods for acquiring assumptions about the user have been reported in the literature [Chi93].

In AVANTI, assumptions will be acquired based on:

1. An initial interview An initial interview allows for the acquisition of primary assumptions about the user and is therefore a valuable source of information for initially assigning the user to certain user subgroups (see 'stereotypes' below).
2. Observing the user's interaction with the system Certain dialog actions performed by the user can be used for the acquisition of primary assumptions. For instance, if the user requests an explanation for a technical term, then it could be assumed that she/he is not familiar with it [KMN94].
3. Inferences Based on primary assumptions about the user and additional information about the application domain, the system can draw inferences in order to acquire further assumptions about the user. For instance, if the user is interested in paintings and, being a tourist, has a special interest in one of the most famous object d'arts of the Louvre, we can anticipate the user's interest in 'The Mona Lisa'.
4. Stereotypes So-called 'stereotypes' [Ri79] contain assumptions about interesting characteristics of user subgroups (e.g., tourists, blind users). If certain preconditions are satisfied, a stereotype can be activated for a specific user which means that the assumptions contained in the stereotype are assigned to the user.

Technical Approach

In AVANTI, the available information will be presented to the user in a hypermedia form. HTML will be the basis for the markup language used for incorporating multimedia objects which are stored in various databases. To adapt the content of the hypermedia document according to the user's presumed knowledge, interests and abilities, a description of a hypermedia page will contain optional and alternative elements. An optional element could be, for instance, additional information on wheelchair accessibility. An alternative element could be a technical vs. a non-technical description or a picture of a painting vs. its verbal description. Based on the assumptions stored in the user model and the aforementioned information related to the semantics of multimedia objects (e.g., explanation, overview, detail), the system presents an optional element if it seems to be useful and chooses the most appropriate one among several alternatives. Such adaptations performed by the system are visible and controlled by the user. This enables him/her to reset any adaptation done by the system, if not desired or appropriate. Therefore, a patronization of the user by the system is avoided.

The general architecture of AVANTI is based on a distributed environment. The different system components cooperate like standard World-Wide Web software (i.e., they are compliant to existing standards like HTTP). A central user modeling server will be developed that accepts information based on the initial interview and the user's actions reported from the user interface. It will maintain stereotypes which contain assumptions about user subgroups and will host inference rules which allow for the acquisition of further assumptions based on existing ones. Available tools in the field of user modeling shell systems (e.g. BGP-MS [KP95]) will be investigated and the most appropriate one will be exploited for the development of the user model server.

References

[Chi93] Chin, D.N. (1993): Acquiring User Models. In: Artificial Intelligence Review 7, 185-197.

[KMN94] Kobsa, A.; Mueller, D.; Nill, A. (1994): KN-AHS: An Adaptive Hypertext Client of the User Modeling System BGP-MS. Proceedings of the Fourth International Conference on User Modeling, Hyannis, MA, 99-105.

[KP95] Kobsa, A.; Pohl, W. (1995): The User Modeling Shell System BGP-MS. User Modeling and User-Adapted Interaction 4(2), 59-106.

[Opp94] Oppermann, R. (Ed.) (1994): Adaptive User Support - Ergonomic Design of Manually and Automatically Adaptable Software. Lawrence Erlbaum Associates, Inc., Publishers, Hillsdale, New Jersey.

[Ri79] Rich, E. (1979): User Modeling via Stereotypes. Cognitive Science, 3:329-354.