Baasye, K., Dean, T., and Kaelbling, L. (1995). Learning dynamics: System identification for perceptually challenged agents. Artificial Intelligence, 72:139--171.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....or robots) are described with formalisms which lend themselves to an analysis in terms of dynamical systems theory. Such formalisms include di#erential equations (e.g. Jaeger and Christaller, 1998) Large et al. 1997) recurrent neural networks (e.g. Beer, 1995) and stochastic automata (e.g. (Baasye et al. 1995)) The most common kind of formal analyis concerns the description of attractor structures emerging from learning or evolution (e.g. Robertson et al. 1993) Smithers, 1995) In the present article I wish to point out a basic property of dynamical systems based robots, namely, the inherent ....

Baasye, K., Dean, T., and Kaelbling, L. (1995). Learning dynamics: System identification for perceptually challenged agents. Artificial Intelligence, 72:139--171.

.... transparent, and often computationally efficient models of continuous systems [26] 15] among them recurrent neural networks [55] In the programming of mobile robots, they are widely used as learnable memory modules for representing temporal experiences, in particular in navigation [4] [38] A simple example of such a system is given in fig. 1. The figure shows a two state stochastic transition graph, which generates stochastic sequences of a s and b s, as follows. At any time t, where t = 0; 1; 2; the system is either in state a or in state b. If it is in state a, then ....

K. Baasye, Th. Dean, and L.P. Kaelbling. Learning dynamics: System identification for perceptually challenged agents. Artificial Intelligence, 72:139--171, 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC