Jim when we want to define the class of should be justified by something Schmolze argued that if you think of people who work in specific institutions), other than the code implementing a sort of lingua franca for knowledge (2) when a concept definition the system. However, interpreting the representation, you can't be committed depends on the assertional properties two terms efficient and principled as to the difference between terminological of its instances (as with gray elephants, worst-case tractability and soundness and assertional knowledge for example), and (3) when and completeness with respect to the or even between roles and concepts.

Thus far, I believe, describing various approximately 120 copies have been limitations of QSIM. At the risk of distributed. The QSIM program is a being scolded again for "employing research tool, not a product, so any universal truths and unarguable commercial rights are retained, and I facts" in support of my position, I cannot warrant that it is free of bugs. Hall examination of the limitations of University of Texas at Austin one's own work is an invaluable Austin, Texas 78712 guide to further research. Akman observes, correctly, that References QSIM is a purely mathematical formalism for expressing qualitative differential Crawford, J.M., Farquhar, A., and Kuipers, 8. 1590 QPC: A Compiler from equation models of the Phvsical Models into Qualitative Differential world, and not a physical modeling Equations In Pr&eedings of the Thank you for publishing our reply Akman's letter refers to his difficulties to Prof. Kuipers in the last issue.

If classical planners are ever to automatically plan the actions of the smart machines, particularly robots for the automatic assembly of industrial objects, then they will have to know much more about geometry and topology as well as sensing. Consider that the simple act of changing an object's grasp -- the change might be necessitated by the nature of some assembly goal -- involves the interaction of the geometries of the grasping device and the object if the change is to occur without a collision between the device and the object. Of course, one could ask, Could geometric considerations be divorced from the highly developed symbolic-level planning? That is, could we first synthesize a symbolic plan and then plug in the geometry for the execution of the actions? Experience has shown the answer to, unfortunately, be a big no.

High-level robotic assembly planning is concerned with how bodies fit together and how spatial relationships among bodies are established over time. To generate an assembly task specification for robots, it is necessary to represent the geometric shapes of the assembly components in a computational form. One of the principal aspects of shape representation that is relevant for assembly tasks is the symmetry of the shape. Group theory is the standard mathematical tool for describing symmetry. The interaction between algebra and geometry within a group theoretic framework has provided us with a unified computational treatment of reasoning about how parts with multiple contacting features fit together.

A significant problem in designing mobile robot control systems involves coping with the uncertainty that arises in moving about in an unknown or partially unknown environment and relying on noisy or ambiguous sensor data to acquire knowledge about that environment. We describe a control system that chooses what activity to engage in next on the basis of expectations about how the information returned as a result of a given activity will improve its knowledge about the spatial layout of its environment. Certain of the higher-level components of the control system are specified in terms of probabilistic decision models whose output is used to mediate the behavior of lower-level control components responsible for movement and sensing. The control system is capable of directing the behavior of the robot in the exploration and mapping of its environment, while attending to the real-time requirements of navigation and obstacle avoidance.

Yoshihiro Morl Kazuhiko Yokosawa ATR Auditory and Visual Perception Research Laboratories 2-1-61 Shiromi Higashiku Osaka 540 Japan ABSTRACT A neural network is applied to the problem of recognizing Kanji characters. The recognition accuracy was higher than that of conventional methods. An analysis of connection weights showed that trained networks can discern the hierarchical structure of Kanji characters. This strategy of trained networks makes high recognition accuracy possible. Our results suggest that neural networks are very effective for Kanji character recognition. 1 INTRODUCTION Neural networks are applied to recognition tasks in many fields.

The performance of these networks has been better than that of conventional methods.

The performance of these networks has been better than that of conventional methods.

ALVIS is a reinforcement-based connectionist architecture that learns associative maps in continuous multidimensional environments. The discovered locations of positive and negative reinforcements are recorded in "do be" and "don't be" subnetworks, respectively. The outputs of the subnetworks relevant to the current goal are combined and compared with the current location to produce an error vector. This vector is backpropagated through a motor-perceptual mapping network.

A pool of handwritten signatures is used to train a neural network for the task of deciding whether or not a given signature is a forgery. The network is a feedforward net, with a binary image as input. There is a hidden layer, with a single unit output layer. The weights are adjusted according to the backpropagation algorithm. The signatures are entered into a C software program through the use of a Datacopy Electronic Digitizing Camera. The binary signatures are normalized and centered. The performance is examined as a function of the training set and network structure. The best scores are on the order of 2% true signature rejection with 2-4% false signature acceptance.