The performance of online algorithms for learning dichotomies is studied. In online learning, the number of examples P is equivalent to the learning time, since each example is presented only once. The learning curve, or generalization error as a function of P, depends on the schedule at which the learning rate is lowered.

Most neural network control architectures originate from work presented by Narendra[I), Psaltis[2) and Lightbody[3). In these architectures, an identification neural network is trained to function as a model for the plant. Based on the neural network identification model, a neural network controller is trained by backpropagating the error through the identification network. After training, the identification network is replaced by the real plant. As is illustrated in Figure 1, the controller receives external inputs as well as plant state feedback inputs. Training procedures are employed such that the networks approximate feed forward control surfaces that are functions of external inputs and state feedbacks of the plant (or the identification network during training).

In supervised learning, learning from queries rather than from random examples can improve generalization performance significantly. We study the performance of query learning for problems where the student cannot learn the teacher perfectly, which occur frequently in practice. As a prototypical scenario of this kind, we consider a linear perceptron student learning a binary perceptron teacher. Two kinds of queries for maximum information gain, i.e., minimum entropy, are investigated: Minimum student space entropy (MSSE) queries, which are appropriate if the teacher space is unknown, and minimum teacher space entropy (MTSE) queries, which can be used if the teacher space is assumed to be known, but a student of a simpler form has deliberately been chosen. We find that for MSSE queries, the structure of the student space determines the efficacy of query learning, whereas MTSE queries lead to a higher generalization error than random examples, due to a lack of feedback about the progress of the student in the way queries are selected.

Most neural network control architectures originate from work presented by Narendra[I), Psaltis[2) and Lightbody[3). In these architectures, an identification neural network is trained to function as a model for the plant. Based on the neural network identification model, a neural network controller is trained by backpropagating the error through the identification network. After training, the identification network is replaced by the real plant. As is illustrated in Figure 1, the controller receives external inputs as well as plant state feedback inputs. Training procedures are employed such that the networks approximate feed forward control surfaces that are functions of external inputs and state feedbacks of the plant (or the identification network during training).

The performance of online algorithms for learning dichotomies is studied. In online learning, the number of examples P is equivalent to the learning time, since each example is presented only once. The learning curve, or generalization error as a function of P, depends on the schedule at which the learning rate is lowered.

The performance of online algorithms for learning dichotomies is studied. In online learning, thenumber of examples P is equivalent to the learning time, since each example is presented only once. The learning curve, or generalization error as a function of P, depends on the schedule at which the learning rate is lowered. For a target that is a perceptron rule, the learning curve of the perceptron algorithm can decrease as fast as p-1,if the schedule is optimized. If the target is not realizable by a perceptron, the perceptron algorithm does not generally converge to the solution with lowest generalization error.

Most of the recent emphasis in the neural network control field has no error feedback as the control input, which rises the lack of adaptation problem. The integrated architecture in this paper combines feed forward control and error feedback adaptive control using neural networks. The paper reveals the different internal functionality of these two kinds of neural network controllers for certain input styles, e.g., state feedback and error feedback. With error feedback, neural network controllers learn the slopes or the gains with respect to the error feedback, producing an error driven adaptive control systems. The results demonstrate that the two kinds of control scheme can be combined to realize their individual advantages. Testing with disturbances added to the plant shows good tracking and adaptation with the integrated neural control architecture.

In supervised learning, learning from queries rather than from random examples can improve generalization performance significantly. Westudy the performance of query learning for problems where the student cannot learn the teacher perfectly, which occur frequently in practice. As a prototypical scenario of this kind, we consider a linear perceptron student learning a binary perceptron teacher. Two kinds of queries for maximum information gain, i.e., minimum entropy, are investigated: Minimum student space entropy (MSSE)queries, which are appropriate if the teacher space is unknown, and minimum teacher space entropy (MTSE) queries, which can be used if the teacher space is assumed to be known, but a student of a simpler form has deliberately been chosen. We find that for MSSE queries, the structure of the student space determines theefficacy of query learning, whereas MTSE queries lead to a higher generalization error than random examples, due to a lack of feedback about the progress of the student in the way queries are selected.

This paper presents an approach to learning from situated, interactive tutorial instruction within an ongoing agent. Tutorial instruction is a flexible (and thus powerful) paradigm for teaching tasks because it allows an instructor to communicate whatever types of knowledge an agent might need in whatever situations might arise. To support this flexibility, however, the agent must be able to learn multiple kinds of knowledge from a broad range of instructional interactions. Our approach, called situated explanation, achieves such learning through a combination of analytic and inductive techniques. It combines a form of explanation-based learning that is situated for each instruction with a full suite of contextually guided responses to incomplete explanations. The approach is implemented in an agent called Instructo-Soar that learns hierarchies of new tasks and other domain knowledge from interactive natural language instructions. Instructo-Soar meets three key requirements of flexible instructability that distinguish it from previous systems: (1) it can take known or unknown commands at any instruction point; (2) it can handle instructions that apply to either its current situation or to a hypothetical situation specified in language (as in, for instance, conditional instructions); and (3) it can learn, from instructions, each class of knowledge it uses to perform tasks.

Analogy has a rich history in Western civilization. Over the centuries, it has become reified in that analogical reasoning has sometimes been regarded as a fundamental cognitive process. In addition, it has become identified with a particular expressive format. The limitations of the modern view are illustrated by monster analogies, which show that analogy need not be regarded as something having a single form, format, or semantics. Analogy clearly does depend on the human ability to create and use well-defined or analytic formats for laying out propositions that express or imply meanings and perceptions. Beyond this dependence, research in cognitive science suggests that analogy relies on a number of genuinely fundamental cognitive capabilities, including semantic flexibility, the perception of resemblances and of distinctions, imagination, and metaphor. Extant symbolic models of analogical reasoning have various sorts of limitation, yet each model presents some important insights and plausible mechanisms. I argue that future efforts could be aimed at integration. This aim would include the incorporation of contextual information, the construction of semantic bases that are dynamic and knowledge rich, and the incorporation of multiple approaches to the problems of inference constraint.