Adaptive Graphical Interfaces
The entire human race represents a great diversity in personality, moods, background, preferences, motivation, goals, education and cognitive skills. Similarly, computers too display variation in purpose, functionality, structure, size and the manner in which their internal functioning are represented. These computer systems have been ultimately designed to be used by human beings and therefore the complexity of the human computer interaction (HCI) is something that must be considered seriously. HCI does not only involve political, organizational and social factors but also the demands of a situation along with user support. The usability of a computer system depends on the user interface displayed on the computer and the human in front of it. The different command names, icons and signs displayed on the screen conveys different meanings for different users and therefore the responses also vary. This increasing complexity of human computer interaction has been the subject of many studies and active efforts are underway to decrease this complexity and increase computer usability through various methods. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Schneider-Hufschmidt, Adaptive User Interfaces Class 8113d, Fall 94)
One way in which the system can be personalized is by customization where the user himself/herself makes certain changes to the system which personalizes the system to their individual needs. These changes are initiated by the user only and depend totally on the level of awareness and knowledge of the computer system in which he/she is working. The other way to make the computer system more usable and the human computer interaction less complex is to make the system itself initiate and execute the personalized and user-centric changes. To do so, the system must be able to obtain some vital information regarding the user by means of some kind of inference mechanism and thus coming up with a user model. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Schneider-Hufschmidt, Adaptive User Interfaces Class 8113d, Fall 94); (Karwowski, 1004)
One of these methods involves the use of “adaptive graphical interfaces” or “adaptive user interfaces.” The objective behind designing an adaptive user interface is to customize the interactive behavior of a system in such a way as to consider both the changing conditions inside an application environment as well as the specific individual requirements of the users. Adaptive user interfaces have the flexibility to change both functionalities and displays corresponding to the user capabilities, needs and preferences by monitoring the user computer interaction. Adaptive interfaces should help users to accomplish their tasks with fewer actions. The eventual goal of adaptive systems is to present an interface that contains only those functionalities, contents and features that the user specifically wants or needs and nothing more. However, the system may also be able to predict certain functions that the user may need in the future. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Schneider-Hufschmidt, Adaptive User Interfaces Class 8113d, Fall 94); (Karwowski, 1004)
Adaptive graphical interfaces can not only improve a user’s performance but also system performance and quality of human computer interaction. Such interfaces can help to get rid of problems arising from information overflow or system complexity. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Schneider-Hufschmidt, Adaptive User Interfaces Class 8113d, Fall 94); (Karwowski, 1004) Adaptive graphical interfaces possess a tremendous amount of potential for providing assistance to a broad range of users operating across a wide span of work contexts. Plenty of research has gone into the development of such systems. Computer systems can be made adaptable if it is provided with an appropriate theory of interaction along with the necessary instructions of how this interaction can be improved. The representations and structure offered at the interface can be made to complement the user’s individual needs, desires and preferences if the computer is arranged to alter its functioning. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Jacko; Sears, 518)
Today, the most common form of adaptive graphical interfaces can be seen in the changes witnessed in menu items displayed in an application wherein the number and type of menu items changes dynamically depending on the most recent choices made by the user. Less accessed menu items cannot be immediately seen in the menu and have to be selected by going through another action. Adaptive interfaces can take on a more complex form when the interfaces change their functionality or display in real time and adjust to the current user’s preferences and needs. Such interfaces have also been referred to as Dynamically Adaptive Interfaces or DAI. For example, in the field of aviation, a computer system displaying Adaptive Automation — AA may display only that information to the pilot which is relevant as well as dependent on the conditions like system state, current workload, etc. existing at that particular point of time. (Karwowski, 1007)
Early research works conducted for developing adaptive GUI — Graphical User Interface include the MERCATOR project conducted by Mynatt and Weber in 1994 and the ‘GUIB or Textual and for Blind People’ project conducted by Petrie, Morley and Weber in 1995. Mercator developed interfaces which modeled graphical components and built up hierarchical relationships between other objects. Both the models could predict user interactions and tried to establish “environment-level adaptations to GUIs” in order to increase their accessibility. Other developments in making user interfaces more adaptable include nonspeech sounds, digitized and synthetic speech and refreshable Braille. Leading developments in nonspeech sound research have resulted in earcons and auditory icons. Earcons, which were developed by Blattner, Sumikawa and Greenberg in 1989, basically utilize nonspeech audio in the graphical user interfaces which provides the user with audio messages about computer operations or objects. Auditory icons, which were developed by Gaver in 1989, include everyday sounds that occur all round us and are mapped in relevant fashion in the computer system. (Jacko; Sears, 524)
Adaptive GUIs are not meant only for the visually or physically impaired but also for infrequent or novice users who can gain from interface dialogue styles which facilitate recognition of commands and through fill-in form dialogue styles and menu choices which serve to decrease their memory load. Adaptive user interfaces should be attuned to the needs of the more experienced user who may simply get irritated by the “overly helpful” icons, sounds and other such objects. The user interface should be able to adapt to such experienced users as well by providing command interfaces or other such interfaces where the user does not feel restricted. (Stephanidis; Jacko, 385)
According to Rothrock and colleagues, the design and modeling of adaptive interface systems can follow two different approaches. The first approach called the “human factors approach” should tackle adaptation from the perspective of automation. Here, the main focus should be on a suitable level of automation — LOA and the splitting up of the tasks between the user and the computer system. This refers to the decisions regarding what kind of tasks should be automated and what should be left to the user, at what point automation should be enables or disabled and who would do the task of enabling or disabling of the automation system. The human factors approach also includes the measurement and analysis of the user’s mental workload and cognitive ability and resources. The second approach outlined by Rothrock et al. is the HCI or “human computer interaction” approach. This involves research in mechanisms that help the system to adapt to the user, situation or task dynamically. In this context, “intelligent user interface” is often used. (Karwowski, 1010)
A user interface may be referred to as “intelligent” in the degree to which it makes adaptations, dynamically takes communication decisions and adapts itself at run-time. According to Maybury, intelligent user interfaces help in establishing a human computer interaction which is more natural trying to imitate the communication that occurs between two humans. In the context of adaptive user interfaces, researchers have identified a few variables which are required for the system to adapt to the user. These variables include user knowledge, user situation awareness, user goals, user performance, task variables like system variables and situation variables, user workload, user personality, user cognitive style and groups of users. (Evangelos Triantafillou, Elissavet Georgiadou, 357); (Stephanidis, 391)
Some other configurable or adaptive interface designs that have been deeply researched include the extremely flexible abstract widgets. Abstract widgets refer to a modeling approach which uses a “semantic abstraction of user interaction.” The separation of the application functionality from the user interface in this modeling approach facilitates users to interact with whichever interface they select, irrespective of their environment.
PAT or Pervasive Accessible Technology is yet another approach which helps disabled individuals to make use of the usual interface devices to interact with the computer system infrastructures. Depending on the disability of the user the UIMS or User Interface Management System model offers the required versatility to adapt the user interface in accordance to their needs. Research has also been carried out on hypermedia systems concentrating on development of adaptive applications which monitor the evolving aspects of the user like domain knowledge and preferences. A user model is created with the help of this information and this model is in turn used as a basis for establishing the user interface adaptation. Task models have also been used for designing adaptive hypermedia. Different types of computer users can be associated with different task models. Task models are used to depict the activities that are to be performed from the user’s point-of-view. (Jacko; Sears, 518)
Adaptive navigation support for hypermedia systems has also been explored as a means of personalizing or adapting user interface. Several prototypes have also been developed to show the way different navigational possibilities can be presented on the basis of user models. In recent times, research has concentrated on the mechanism of abstractions of objects to produce “operationally reliable software infrastructures” that provide alternative physical realizations. Development of systems like JavaBeans by SunSoft and ActiveX by Microsoft representing componentized technology shows the efforts undertaken by the mainstream division of the it industry to offer technological structures to provide accessibility support and more adaptive interfaces. (Jacko; Sears, 520) Another recent development in this field has been the development of the AVANTI system — a “single unified user interface” which employs a rule-based adaptivity technique called and has been designed for disabled users. (Aykin, 207)
Adaptive graphical interfaces have a broad range of applications in a wide variety of fields. The field of medical information is one such domain where adaptive graphical interfaces through hypermedia navigation can be made possible. Hypermedia navigation permits relevant information to be directly accessed and the disposal and data presentation mode can be adapted based on the user needs. A clinical workstation prototype, called HEMA — Health data Manager Application has been proposed by French researchers. The interface provided by this prototype is able to take psychological behavior and cognitive characteristics of each user into account. This prototype can capture and implement a user’s manner of working as well as his or her knowledge. The navigational links in this proposed hypermedia application contains both generic and specific links. Generic links are inbuilt into the system based on the knowledge gathered by domain experts and help to create pertinent links which can be instantiated by the system itself during consultation of patient records. On the other hand, specific links can be set up by the user while consulting patient records and can be established to create a new relationship between two areas of information. (Patel; Rogers; Haux, 132)
Another design of an adaptive graphical interface has been put forward by Igarashi et al. For teleoperation system. Teleoperation systems require a fine balance of human abilities as well as computer processing. An adaptive GUI that presents appropriate displays to increase an operator’s efficiency and reduce errors by adapting to his/her level of knowledge/cognition would be a significant advancement in this field. In teleoperation, communicational constraints result in a less than adequate supply of feedback information with regard to quantity and quality. (Igarashi; Takeya; Kubo; Suzuki; Harashima; Kakikura; Industrial Electronics Society, 6)
Another problem is that the cognition capabilities of humans and the display functions of the user interfaces possess many limitations even though it is expected that the human operators comprehend a significant portion of the feedback information. On the other hand, it is also a fact that the ability of human beings to create a global plan in response to a potential problem contributes to intelligent operations and a degree of flexibility that no autonomous robots can accomplish. This adaptive GUI design proposed an “effective alert function” based on the characteristics of human cognition which would reduce misrecognition on part of the human operator. These human cognition characteristics may vary based on the position and information emphasis media. Additionally, this design also measures the human sensitivities to the GUI so as to ensure an effective and alert adaptive user interface. (Igarashi; Takeya; Kubo; Suzuki; Harashima; Kakikura; Industrial Electronics Society, 6)
A significant challenge in the field of adaptive GUI development is the ability to create realistic adaptivity which actually enhances the human computer interaction and increases the usability of the system. Nothing can be gained from a costly adaptivity system which provides a minimal enhancement in usability. It would be of no use to create an adaptive interface in response to a user characteristic if that particular characteristic cannot be discreetly or reliably deduced from the system-user interaction. Wrong inferences may reduce the predictability of the system and the confidence of the users in the accuracy and functioning of the system may decrease. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Karwowski, 987)
It is ironical that most of the problems related to system usability arise due to the flexible features of the adaptive interfaces. In fact, flexibility and usability are inversely proportional to each other and one of the main challenges of designing an effective adaptive interface is to find the optimal combination of these two important system characteristics. Lack of consistency poses another problem with regard to usability. According to research carried out in the field of human computer interaction, it has been seen that user performance is supported by user interfaces that display consistency. On the other hand, the interactive nature of adaptive user interfaces leads to lack of consistency which may eventually lead to poor system performance. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Karwowski, 988)
An adaptive graphical interface may be considered effective if the quality of the information provided by the system is high, accurate, and has an optimal performance time. In addition, its effectiveness also depends on the subjective evaluations of the users. There is a serious dearth of studies regarding the advantages of adaptive systems over non-adaptive ones and the evaluation of adaptive interfaces must deal with two particularly important problems. Firstly, the outcomes, both positive and negative of newly introduced adaptations to graphical user interfaces may not appear immediately but after extensive and more or less long-time use. Research work involving short-term observations or experiments may not provide the correct representation of the actual scenario. Secondly, adaptivity may result in dynamic changes in the properties or functionalities of the computer system which may have differential effects on various users. Some users may find some interface adaptations to be advantageous whereas others may find the same adaptations confusing and would find it better to have them switched off. If adaptive graphical user interfaces have to achieve success, then these potential variations in user reactions must be considered. (Benyon, Accommodating Individual Differences through an Adaptive User Interface); (Karwowski, 988)
An important point in the context of adaptive graphical interfaces is the use of culturally adaptive interfaces. There has not been much research in this area since the growth of Internet has led to most of the work being conducted in the adaptive hypermedia systems or web-dependent. As a result, studies in other areas have more or less stagnated. An adaptation to the user’s cultural frame is not something that the global user is accustomed to and is more familiar with software that embeds the cultural values of the Western world. On the other hand, culturally adaptive interfaces may be distinctly helpful and may even increase the efficiency of the user. It may be possible to map specific cultural behavior onto a user interface model. This would require identification of universally applicable cultural markers that are valid across all user interfaces. (Aykin, 210)
Again, the usability factor must be thoroughly investigated in every case. In this context, it is important to remember that the average global user is not accustomed to having cultural adaptations made to his/her graphical interface and may initially feel uncomfortable and even reject the adaptations. Repeated and prolonged use may overcome user resistance and the advantages may finally outweigh the uneasiness of dealing with the new modifications. In these cases where adaptations are extremely important but infrequent, it is better to choose computer-supported adaptation, where control lies more with the user, over automatic adaptation. A useful addition to such an adaptive GUI would be to incorporate an easily accessible history log which tracks the recent adaptive changes which also permits the user to retract changes at any point of time. (Aykin, 210)
In recent times, there has been a proposal to develop adaptive and self-managed GUI using agent-based technology. Such an approach may become necessary since this technology allows an expansion or modification of the system to be anticipated and implemented during run-time. It also allows certain components of the GUI to have autonomous features. Such adaptive GUIs may possess various diverse or complex kinds of interaction between internal components along with interaction with heterogeneous resources that are externally distributed. This technology also has the advantage of providing fault tolerance. (Kernchen; Dumke, 40)
Having discussed a wide variety of technologies and techniques being pursued in the development of adaptive graphical interfaces, one can understand the significance of having adaptive mechanisms inbuilt into the system as well as providing some degree of control to the user. Another factor which has emerged from the discussion is the usability criterion which cannot be ignored as the entire success of an adaptive technology rests on it.
References
Aykin, Nuray. Usability and Internationalization: Global and local user interfaces.
Springer. 2007.
Benyon, David. Accommodating Individual Differences through an Adaptive User
Interface.
Evangelos Triantafillou, Elissavet Georgiadou, Anastasios a. Economides, Applying adaptive variables in computerised adaptive testing. Australasian Journal of Educational Technology, 2007, vol. 2, no. 3, pp: 350-370.
Igarashi, H; Takeya, a; Kubo, Y; Suzuki, S; Harashima, F; Kakikura, M; Industrial
Electronics Society. Human adaptive GUI design for teleoperation system. 31st Annual Conference of IEEE, vol: 6-10, Nov. 2005, p. 6.
Jacko, Julie a; Sears, Andrew. The human-computer interaction handbook. Lawrence Erlbaum Associates. 2003.
Karwowski, Waldemar. International encyclopedia of ergonomics and human factors.
CRC Press. 2006.
Kernchen, Steffen; Dumke, Reiner R. Developing Adaptive and Self-Managed Graphical
User Interfaces. IEEE Mult-disciplinary Engineering Education Magazine, vol. 2, no. 2, June 2007. pp: 38-43.
Patel, Vimla L; Rogers, Ray; Haux, R. Medinfo 2001: Proceedings of the 10th World
Congress on Medical Informatics (Studies in Health Technology and Informatics, 84). IOS Press. 2001.
Schneider-Hufschmidt, Matthias. Adaptive User Interfaces Class 8113d, Fall 94. 1994.
Stephanidis, Constantine; Jacko, Julie a. Human-Computer Interaction: Theory and Practice (part 2), Volume 2 (Human Factors and Ergonomics). CRC. 2003.
