Agent computing themes in biologically inspired models of learning and development

Abstract

After evaluating general features and attributes of the agent notion, the overlap of features in candidate (attribute) cores, and several less central features, the paper addresses agent and related theory in neuroscience, observing how agent notions have penetrated portions of this field and how the field itself emphasizes and further develops some agent themes via, e.g. schema theory, neural net-artificial intelligence (AI) comparisons, and other research. In remaining sections, models for development of memory strategies in children are presented, illustrating cooperative and competitive neural modeling agents, an active role for a "human agent in the loop," and integrating broadly-based neural network (NN) modeling with other bio-inspired models.

Applied Research in Uncertainty Modeling and Analysis

“…The prior knowledge of both "unchanged" and "changed regions was modeled by energy functions, which represented the potential of a pixel being in the corresponding status ("unchanged/"changed"). Another approach based on the iterative conditional mode algorithm [21], employed in [18], has a low computational cost, but may converge to a local extremum. We believe this is a powerful framework to model the uncertainty and minimize the error rate in detecting changes.…”

“…We select a neuron c whose output is the largest among the neural network c = arg max y, (k) j (18) If there are some neurons c, 2 0, j = 1, 2, K , that satisfy Eq. (18), then the neuron with the largest sum of weights is selected.…”

“…We evaluated equations (8) through (10) using Maple [18] and then the graph of is shown in Figure 2, without dropping the graph straight down to the x-axis at the end points. We evaluated equations (8) through (10) using Maple [18] and then the graph of is shown in Figure 2, without dropping the graph straight down to the x-axis at the end points.…”

“…We evaluated equation (18) using Maple [18] for 0.01 5 , O 5 1 and the graph of the fuzzy service rate, without dropping the graph straight down to the x-axis at the end points, is in Figure 3. We evaluated equation (18) using Maple [18] for 0.01 5 , O 5 1 and the graph of the fuzzy service rate, without dropping the graph straight down to the x-axis at the end points, is in Figure 3.…”

“…A second is that our current work is a predecessor of some agent style computing [18]- [21], a mode, we interpret, as calling for explicit solution patterns for the "domain" problem (to offset complexities in the networking and brokering features) and to facilitate a set of collaborators to engage in this kind of modeling effort. The computational details are important for more than one reason.…”