Jueves 7 de Septiembre, 14:00 hs. (Aula II)

Dr. Yves Demazeau
Centre National de la Recherche Scientifique
LEIBNIZ, Grenoble, France
 

Merging between the physical world and the virtual world

A couple of years ago, just before the explosion of the World Wide Web, much hype was created about Virtual Reality (VR). Now, some years later, the hype has passed (and has been replaced by other hype), but the importance of VR has not changed. The 3-D capabilities of personal computers are steadily improving and the World Wide Web is counting an increasing number of sites that feature content in VRML or competing languages. Today's chat systems are evolving into VR systems and already feature VR elements like avatars. Multi-user games, pushing the envelope in consumer VR, are gaining popularity every day. It is conceivable that chat systems and multi-user games will essentially merge in a not so distant future. Information in general on the Web will most often be presented in a VR format. Similarly, virtual agents will be increasingly present in the physical world. Animal-like and human-like robots, connected to the network, will interact with humans (and each other) in a physical way, although they essentially exist in the virtual world. We might call this technology Real Virtuality (RV). Already toys exist that are based on this philosophy. Some work on humanoid robots also fits in quite well. How to incorporate a social dimension in VR and RV is basically an unanswered question and the research opportunities in this context are plenty.

Mutual acceptance/tolerance between human and artificial agents

We can expect artificial agents to take on an increasingly important role in the multi-user (multi-agent) systems described above. For example, they might intervene in conversations between people (in a virtual room) to offer information on the topic of the conversation or even on the social context in which the conversation is held. Humans adapt quite well to changing environments and we can expect them to adapt quickly to these agents that play a peripheral, assisting role. If these agents obey certain rules, they might be taken for granted quite easily. In certain contexts, one might not even care whether one is dealing with a real or artificial agent. Actually, there already exist examples of artificial agents on News groups that were taken for human agents. Artificial agents, on the other hand, are having a hard time understanding our physical world, since it is so different from theirs. Our social world, on the other hand, might make much more sense to them since they will live in a social world themselves: the isolated agents of today will have no choice but to socially organise themselves. It is this correspondence between social worlds that will help to bridge the gap between natural and artificial agents. There exist some work here and there, but in general this domain is not much explored, and much work has still to be done.

High transparency of HCI towards a unique interaction language

A consequence of the evolution described in the two previous sections is that human-computer interfaces will become increasingly transparent. As far as the interaction medium goes, the evolution from text to images, sound and other more natural media will continue. How to represent agents in a multi-media way will be an important problem to investigate. Similarly, how to represent ourselves to artificial agents is another question that has to be addressed. Focusing on interaction : much work has been in Natural Language Processing; Man-Machine Interfaces have been developed trying to reach the level of natural language; and Interaction Languages between artificial agents like KQML or ACL are usually based on Speech Act theory. As soon as the merging between the physical world and the virtual world is established, as soon as human and artificial agents will accept or at least tolerate each other, one should expect the arising of a unique interaction language between agents. It is difficult to predict how this language will look like but it will certainly subsume the current interests of the three kinds of languages: artificial / artificial, artificial / human, human / human. Figuring out the semantics and pragmatics of such a language will be a key issue in the coming years.

Computer-assisted management of social space

Much technological advance in the history of mankind has been focused on the management of physical space. Buildings not only provide shelter, but they also organise physical space. Increasingly sophisticated transportation technology serves to overcome the constraints of physical space. Until recently, social space was very much related to physical space. To interact with somebody, one had to be physically close. Technology has changed all this. Telephony, for example, allows us to interact with people far away in physical space. Until the advent of the widespread use of the Internet, however, these interactions were mostly of the one-to-one or the broadcast type. People did not fundamentally change their social space through technology. Internet chat rooms, to give one example, have changed all this. People can now make friends and enemies who are on the other side of the world, without meeting in physical space. Social space has to be organised, though, and right now, on the Internet and in other domains, this is done in an ad hoc fashion. In short, there is no dynamic management of social space as it exists in physical space where people can move to get closer to or farther away from other people. How to manage social space dynamically is a question that will have to be addressed in the future.

Evolution from productivity tools to creativity tools

The advances outlined in the four preceding sections will greatly simplify the production of material, whether physical or virtual. As barriers disappear, the need for middlemen will gradually disappear. Take, for example, software production. Investors, managers, programmers and graphical artists can easily find themselves and interact through the automatic management of social space and the transparency of interfaces. Using similar social networks, the product finds its way to the consumers. Pricing can be partly automatic as a side-effect of the process. As production becomes smoother, the importance of creativity will increase. We will see an evolution from productivity tools to creativity tools. The creation of original content will become one of the main activities in hi-tech industry. Workers will want tools to assist themselves in this process, to support their creativity. In film-making, for example, people will build quick prototypes, using virtual actors to try out ideas. As we have now a market in real actors, a market in virtual actors will emerge. Virtual actors are just one example. Virtual musicians are another, focusing on a single medium but being of equal importance, offering an adequate framework to investigate such issues. How to move from productivity to creativity constitutes an important field of research.

Increased focus on real-time and highly interactive domains

Passive information is devaluating quickly. People want information to be interactive, to follow them wherever they want to go. Information will not only have to be increasingly multi-media and original, it will have to adapt to its consumers in a real time fashion. Let us take an example from the music business. From a consumer point of view, static recordings will no longer do. Consumers will want a live experience whenever possible. From a producer point of view, static recordings will become less and less interesting due to the increasing ease of copying and distribution through the Internet. The producers will have to return to the live production of their music. As many of them are relying on extensive studio work, this will only be possible if they will be assisted by interactive real-time agent-like technology, i.e. virtual musicians. Whether virtual musicians will merely accompany human musicians or constitute a complete virtual band, they will play an important role in tomorrow's music business. Another key example illustrating this perspective is Intelligent Manufacturing Systems which are highly flexible, interactive, real-time, and are becoming agent-based. How to construct interactive real-time agent-like technology, and how to let agents interact with their tutors and their peers are basic questions that will have to be addressed in the near future.

Notes