Specification of intelligent virtual agents behavior by demonstration / Especificação do comportamento de agentes virtuais inteligentes por demonstração
AUTOR(ES)
Gênesis Barros Campos
DATA DE PUBLICAÇÃO
2009
RESUMO
The possibility to test hypotheses and to experience reality without the expenses and dangers of the "real world" have attracted the attention of professionals to the use of virtual environments. As importantas interacting in realistic environments is dealing with characters that convey a sense of living, "intelligent" beings. The definition of the layer of action selection is crucial to the realism of the character and it can be identified from the examples given by the users of what the agent must do.Faced with the existent end-user programming approaches not exploring this feature, this paper presents a new proposal for the specification of intelligent virtual agents. The proposed solution incorporates strengths of parametric, descriptive and imitative programming paradigms, and facilitates the development of agents to those who would like to do it but are not able because of lack of programming knowledge. A system was developed to validate the possibility of specifying the agents behavior using the new end-user programming approach. The environment developed incorporates the realistic visuals of current virtual environments, and by itself, motivates both developers and users. The results achieved at this degree of complexity suggest that the new approach can be applied to a wide range of situations, and that it can spread the use of this resource.
ASSUNTO(S)
ciencia da computacao agentes virtuais inteligentes demonstration programming programação por usuários finais intelligent virtual agents end-user programming programação por demonstração
Documentos Relacionados
- TRUST IN INTELLIGENT AGENTS
- AGENTES INTELIGENTES PARA DETECÇÃO DE INTRUSOS EM REDES DE COMPUTADORES
- Accounting information systems: modeling and intelligent agents application
- HIERARQUICAL NEURO-FUZZY MODELS BASED ON REINFORCEMENT LEARNING FOR INTELLIGENT AGENTS
- Semiotic analysis of neuroenergetic networks in the construction of intelligent agents