Ferramentas do usuário

Ferramentas do site


geopro:pedro:situated

Diferenças

Aqui você vê as diferenças entre duas revisões dessa página.

Link para esta página de comparações

Ambos lados da revisão anteriorRevisão anterior
geopro:pedro:situated [2007/07/06 15:23] pedrogeopro:pedro:situated [2007/07/31 01:57] (atual) pedro
Linha 1: Linha 1:
 ======A Situated View of Representation and Control====== ======A Situated View of Representation and Control======
-|Rosenschein and Kaelbling1995| Artificial Intelligence| [[http://leg.ufpr.br/~pedro/papers/rosenschein95situated.pdf|pdf]]| [[http://scholar.google.com.br/scholar?hl=pt-BR&lr=&cites=17385910363361766699| 94 citations in Scholar]]|+|Rosenschein and Kaelbling1995| Artificial Intelligence| [[http://leg.ufpr.br/~pedro/papers/rosenschein95situated.pdf|pdf]]| [[http://scholar.google.com.br/scholar?hl=pt-BR&lr=&cites=17385910363361766699| 94 citations in Scholar]]| 
 \\ \\
 +
 +**Abstract:** Intelligent agents are systems that have a complex, ongoing interaction with an environment that is
 +dynamic and imperfectly predictable. Agents are typically difficult to program because the correctness of a program depend
 +on details of how the agent is situated in its environment. In this paper, we present a methodology for the design of 
 +situated agents that is based on situated automata theory. This approach allows designers to describe the informational 
 +content of an agent's computational states in a semantically rigorous way without requiring a commitment to conventional 
 +runtime symbolic processing. We start by outlining this situated view of representation, then show how it contributes to 
 +design methodologies for building systems that track perceptual conditions and take purposeful actions in their environments.
  
 {{  http://leg.ufpr.br/~pedro/figures/interaction_agent_environment.jpg?600|Interaction between agent and environment}} {{  http://leg.ufpr.br/~pedro/figures/interaction_agent_environment.jpg?600|Interaction between agent and environment}}
 +
 +\\
  
 Situated agents are very difficult to model because they have close interactions with the environment Situated agents are very difficult to model because they have close interactions with the environment
-they belong. //The emphasis on an agent's connection to its environment is an important change from that +they belong. The emphasis on an agent's connection to its environment is an important change from that 
-of traditional theories of representation and control.// The methodology is based on the //situated automata +of traditional theories of representation and control.  
-theory//. //knowledge is an effective way of describing the relationship between agent and environment//+Knowledge is an effective way of describing the relationship between agent and environment.
  
-//  The important aspect of situated automata theory is modeling systems such that, for each +__Comment:__ Modelling is not so difficult as situated agents in real world. 
 +//The important aspect of situated automata theory is modeling systems such that, for each 
 state of the environment E, there will be a corresponding state of the automaton M (by Carneiro).// state of the environment E, there will be a corresponding state of the automaton M (by Carneiro).//
 I understood that a TerraME agent is situated because of the execution of the jump conditions until I understood that a TerraME agent is situated because of the execution of the jump conditions until
Linha 17: Linha 29:
 the external world.// the external world.//
  
-//We can construct arbitrarily complex machines from machines of two primitive types: pure functions and delay. + 
-Pure functions consists of two locations and specify the values of one location as a function of the other. +Let the environment be represented as a nondeterministic automaton <S, P, A, //init//, v, //out//>, where 
-Delay machines also consists of two locations, but constrain the values of the output location to be the values +  * S is a finite set of states of the environment; 
-that the input location had on the previous tick.//+  * P is a finite set of outputs (there are usefully viewed as percepts from the agent's perspective); 
 +  * A is finite set of actions that the agent can generate as input to the environment; 
 +  * //init// is a set of states containing the one that the environment is known to be in initially; 
 +  * v is a relation on S x A x S where v(s_1, a, s_2) holds if it is possible for the world to make a transition from state s_1 to state s_2 when action a is generated by the agent 
 +  * //out// is a function mapping S to P 
 + 
 + 
 +{{http://leg.ufpr.br/~pedro/figures/situated.jpg|Interaction between agent and environment  }} 
  
geopro/pedro/situated.1183735431.txt.gz · Última modificação: 2007/07/06 15:23 por pedro