Abstract: A data processing includes an on-line, interactive, intelligent help system which provides suggestions as to actions a user can take after entry into the system of an erroneous command or a question. The system also responds with explanations of why the suggestions were made and how they work. The system includes a natural language analyzer for converting the questions into goals. A knowledge base and an inference engine further analyze the goals and provide one or more suggestions on how to achieve such goals. An explanation generator uses such analysis to dynamically generate the explanations which are tailored to the user's goal.

Abstract: A self-repairing trellis network includes a plurality of stages (11, 12 and 13) wherein a first set of N processing elements and a second set of M processing elements are situated. The N processing elements are normally active while the M processing elements are designated as spares. The stages are connected together by paths through variable weighing circuits which are legal or illegal. In response to a combination of input bits, an arrangement maps bit groupings onto the stages to turn on selective cells which form nodes coupled by paths. Only certain paths (legal paths) are tolerated in accordance with convolution coding of the input bit combination. If an error is present, an illegal path attempts to form from the legal path which starts a process of error detection and correction wherein a spare cell may be substituted for a failed cell. Since the substitution serves to maintain the redundancy status of the network, subsequent failures will be detected and corrected.

Abstract: An intelligent help system which processes information specific to a user and a system state is described. The system incorporates a monitoring device to determine which events to store as data in an historical queue. These data, as well as non-historical data (e.g., system state), are stored in a knowledge base. An inference engine tests rules against the knowledge base data, thereby providing a help tag. A display engine links the help tag with an appropriate solution tag to provide help text for display.

Abstract: This is a fully parallel analog backpropagation learning processor which comprises a plurality of programmable resistive memory elements serving as synapse connections whose values can be weighted during learning with buffer amplifiers, summing circuits, and sample-and-hold circuits arranged in a plurality of neuron layers in accordance with delta-backpropagation algorithms modified so as to control weight changes due to circuit drift.

Abstract: A neural network structure includes input units for receiving input data, and a plurality of neural networks connected in parallel and connected to the input units. The plurality of neural networks learn in turn correspondence between the input data and teacher data so that the difference between the input data and the teacher becomes small. The neural network structure further includes output units connected to the plurality of neural networks, for outputting a result of learning on the basis of the results of learning in the plurality of neural networks.

Abstract: A robotic surgical system (10) includes a multiple degree of freedom manipulator arm (14) having a surgical tool (22). The arm is coupled to a controller (24) for controllably positioning the surgical tool within a three dimensional coordinate system. The system further includes a safety monitoring processor (38) for determining the position of the surgical tool in the three dimensional coordinate system relative to a volumetric model. The volumetric model may be represented as a constructive solid geometry (CSG) tree data structure. The system further includes an optical tracking camera system (28, 32) disposed for imaging a region of space that includes at least a portion of the manipulator arm. An output of the camera system is coupled to the processor (38) that processes the volumetric model for determining if the surgical tool is positioned outside of the volumetric model.

Abstract: A work forming apparatus of this invention includes a robot which operates within a predetermined operational range, a working machine with a chuck, a supply loop, a discharge loop, supply pallets for accommodating material works, and discharge pallets for accommodating formed works. An information card is attached to each pallet so as to record work information relative to the corresponding pallet. Two information processing heads are provided for effecting reading and writing of work information at a work supplying section and a work discharging section, respectively. A control device provided for controlling all the above elements inputs a production command so that a specified hand is automatically selected and the robot employs the hand to take the required material works alone out of the supply pallets which are being circulated on the supply loop. The material works are then formed and discharged by the discharge pallet.

Abstract: A multi-layer perceptron circuit device using integrated configuration which is capable of incorporating self-learning function and which is easily extendable. The device includes: at least one synapse blocks containing: a plurality of synapses for performing weight calculation on input signals to obtain output signals, which are arranged in planar array defined by a first and a second directions; input signal lines for transmitting the input signals to the synapses, arranged along the first direction; and output signal lines for transmitting the output signal from the synapses, arranged along the second direction not identical to the first direction; at least one input neuron blocks containing a plurality of neurons to be connected with the input signal lines; and at least one output neuron blocks containing a plurality of neurons to be connected with the output signal lines.

Abstract: A technique for preventing a stalled linear motor device from damage including monitoring accelerometer servos which generate signals representative of the difference between a command acceleration signal controlling the linear motor and the actual acceleration experienced by the linear motor device which signal represents the acceleration error and the direction of the error. The different signal is compared against programmed limits to develop a break-loose signal when the programmed limit (positive or negative) is exceeded. The break-loose signal is monitored as to duration, which duration is preprogrammed according to the dynamics of the system specifications presently in use.

Abstract: In a knowledge data management apparatus, the impact of the methods for acquiring knowledge data are digitized as storage impact numbers which represent the importance degree of being stored, the storage impact numbers are stored together with the knowledge data, the storage impact numbers are changed in accordance with the elapsed storage time, frequency of use and object in which the knowledge data is used, and the knowledge data having a smaller storage impact number than a predetermined value on the basis of a manageable amount of the knowledge data are canceled.

Abstract: A variable weight optical interconnector is disclosed to include a projecting device and an interconnection weighting device remote from the projecting device. The projecting device projects a distribution of interconnecting light beams when illuminated by a spatially-modulated light pattern. The weighting device includes a photosensitive screen provided in optical alignment with the projecting device to independently control the intensity of each projected interconnecting beam to thereby assign an interconnection weight to each such beam. Further in accordance with the present invention, a highly-interconnected optical neural network having learning capability is disclosed as including a spatial light modulator, a detecting device, an interconnector according to the present invention, and a device responsive to detection signals generated by the detecting device to modify the interconnection weights assigned by the photosensitive screen of the interconnector.

Abstract: A random access memory (RAM) circuit is provided wherein an input signal matrix forming an identifiable original pattern is learned and stored such that a distorted facsimile thereof may be applied to generate an output signal matrix forming a replication of the original pattern having improved recognizable features over the distorted facsimile. The input signal matrix is logically divided into a plurality of predetermined subsets comprising a unique element of the input signal matrix and the elements in the neighborhood thereof. Each predetermined subset is quantized into a first digital address and applied at the address inputs of a memory circuit for retrieving data stored in the addressed memory location, while one signal of the predetermined subset is digitized and weighted and combined with the data retrieved from the addressed memory location for storage in the same addressed memory location.

Abstract: The invention is directed to an automated, real-time, rule-based reasoning with uncertainty system and method for practicing same. The system has a three layer structure comprising of representation, inference and control layers. Rule firings and conclusions are computed in the inference layer in a manner to propagate uncertainty information through the system in accordance with predetermined formulas designated by the control layer. Rule structures and conclusions and uncertainty information associated therewith are represented to the user in the representation layer. The control layer includes a planning mechanism which determines the fastest plan for reaching the conclusion and generates prioritized alternate plans which, if time permits, will be executed in accordance with their respective priorities.

Abstract: A fuzzy data comparator receives a fuzzy data digital data bit stream and compares each frame thereof with multiple sets of differing known data stored in a plurality of pattern memories, using a selected comparison metric. The results of the comparisons are accumulated as error values. A first neural postprocessing network ranks error values less than a preselected threshold. A second neural network receives the first neural network solutions and provides an expansion bus for interconnecting to additional comparators.

Abstract: A Japanese language sentence containing a word not registered in an electronic dictionary is divided by following a series of predetermined rules. When more than one division of such a sentence is possible, an evaluation is made in order to determine the best division of the sentence containing a word not registered in the dictionary.

Abstract: Operating a mode change switch changes the mode of a controller to first or third command control mode which is suitable for a work involving a large displacement of a manipulator or a second command control mode suitable for a work requiring precision. When an operation handle of an operation unit is operated, the controller controls the operation unit in each command control mode to drive the manipulator. In a case where a large-displacement-involving work and a precision-requiring work are continuously executed, therefore, the mode of the controller is first changed to the first or third command control mode by operating the mode change switch and then changed to the second command control mode. The manipulator can be driven in the proper command control mode for a work to be done.

Abstract: A computer card game using a video screen which allows the player to make a second wager for an additional card after a first card has been dealt which will be compared with the cards in the first card hand. The matching cards are defined as Jokers or Wild Cards which will then be used to define the best card hand. The amount of the second wager can be fixed and the payout will vary according to the first card hand. Alternatively, the amount of the second wager can vary and the payout will remain the same. The player has the option to terminate the game if he chooses not to register the second wager. The game is won or lost when the card hand as configured is compared to the ranking of card hands on the payout table, which also determines the amount of the payout, if any. In one embodiment, the player is permitted to substitute at least one card in the first card hand before the second wager is registered.

Abstract: A neural network determines optimal control inputs for a linear quadratic discrete-time process at M sampling times, the process being characterized by a quadratic cost function, p state variables, and r control variables. The network includes N=(p+r)M neurons, a distinct neuron being assigned to represent the value of each state variable at each sampling time and a distinct neuron being assigned to represent the value of each control variable at each sampling time. An input bias connected to each neuron has a value determined by the quandratic cost function for the variable represented by the neuron.

Abstract: An industrial robot is able to automatically set the operating conditions thereof depending on a position of installation. The motion of a robot unit is input and controlled by a robot control unit. A first data storage portion of the robot control unit is provided with a storage region preliminarily storing (a) data about installation positions of the robot unit with respect to a reference installation position of the same robot unit, and (b) date about operating conditions with respect to the positions of installation. When data about an installation position upon the use of the robot is input to the robot control unit, the data of the operating conditions, in response to the input data of the position of installation, is read from the first data storage unit, and is automatically set in a second data storage unit to control the motion of the robot unit in accordance with the operating conditions set in the second data storage unit.