Abstract: Apparatus stores a composite video frame including synchronization signals, color burst signals, and analog signals. The synchronization signals, color burst signals and analog signals of a video frame are sampled and digitized by circuitry: Two addressable memories are provided. An input register coupled to first and second memories buffers a plurality of digitized samples. An output register coupled to said first and second memories buffers the outputs of the memories. Address circuits provide sequential addressed to the memories for each cycle. During a write cycle, control means provide control signals for enabling the input register to buffer the samples and alternately write enabling the memories to store said samples. During a read cycle, control signals are provided for alternately read enabling the second memories to output the samples and to enable the output register to buffer the output. A circuit converts the output into analog.

Abstract: A method of training an artificial neural network uses a computer configured as a plurality of interconnected neural units arranged in a layered network including an input layer having a network input, and an output layer having a network output. A neural unit has a first subunit and a second subunit. The first subunit having one or more first inputs, and a corresponding first set of variables for operating upon the first inputs to provide a first output. The first set of variables can change in response to feedback representing differences between desired network outputs for selected network inputs and actual network outputs. The second subunit has a plurality of second inputs, and a corresponding second set of variables for operating upon said second inputs to provide a second output. The second set of variables can change in response to differences between desired network outputs for selected network inputs and actual network outputs.

Abstract: A method of accelerating the training of an artificial neural network uses a computer configured as an artificial neural network with a network input and a network output, and having a plurality of interconnected units arranged in layers including an input layer and an output layer. Each unit has a multiplicity of unit inputs and a set of variables for operating upon a unit inputs to provide a unit output. The computer is programmed with a back propagation algorithm. A plurality of examples are serially provided to the network input and the network output is observed. The examples are iterated and proposed changes to each set of variables are calculated in response to feedback representing differences betwen the network output for each example and the desired output. The proposed changes are accumulated for a predetermined number of iterations, whereupon the accumulated proposed changes are added to the set of variables.

Abstract: A method of accelerating the training of an artificial neural network uses a computer configured as an artificial neural network with a network input and a network output, and having a plurality of interconnected units arranged in layers including an input layer and an output layer. Each unit has a multiplicity of unit inputs and a set of variables for operating upon a unit inputs to provide a unit output. A plurality of examples are serially provided to the network input and the network output is observed. The computer is programmed with a back propagation algorithm for adjusting each set of variables in response to feedback representing differences between the network output for each example and the desired output. The examples are iterated until the signs of the outputs of the units of the output layer converge. Then each set of variables is multiplied by a multiplier. The examples are reiterated until the magnitude of the outputs of the units of the output layer converge.

Abstract: A method of accelerating the training of an artificial neural network uses a computer configured as an artificial neural network with a network input and a network output and having a plurality of interconnected units arranged in layers including an input layer and an output layer. Each unit has a multiplicity of unit inputs and a set of variables for operating upon a unit inputs to provide a unit output in the range between binary 1 and binary 0. A plurality of training examples is serially provided to the network input and the network output is observed. The computer is programmed with a back propagation algorithm for changing each set of variables in response to feedback representing differences between the network output for each example and the desired output. The examples are iterated while the output of a unit is observed.

Abstract: A trainable artificial neural network includes a computer configured as a plurality of interconnected neural units arranged in a layered network. An input layer has a network input and an output layer has a network output. A neural unit has a first subunit and a second subunit, with the first subunit having one or more first inputs and a corresponding first set of variables for operating upon the said first inputs to provide a first output. The first set of variables can change in response to feedback representing differences between desired network outputs and actual network outputs. The second subunit has a plurality second inputs, and a corresponding second set of variables for operating upon said second inputs to provide a second output. The second set of variables can change in response to differences between desired network outputs for selected network inputs and actual network outputs.

Abstract: A method of accelerating the training of an artificial neural network uses a computer configured as an artificial neural network with a network input and a network output, and having a plurality of interconnected units arranged in layers including an input layer and an output layer. Each unit has a multiplicity of unit inputs and a set of variables for operating upon a unit inputs to provide a unit output in the range positive 1 and negative 1. A plurality of examples are serially provided to the network input and the network output is observed. The computer is programmed with a back propagation algorithm for calculating changes to the sets of variables in response to feedback representing differences between the network output for each example and the desired output. The absolute magnitude of the product of an input and the corresponding output of a unit is calculated.

Abstract: A method of training an artificial neural network uses a first computer configured as a plurality of interconnected neural units arranged in a network. A neural unit has a first subunit and a second subunit. The first subunit has first inputs and a corresponding first set of variables for operating upon the first inputs to provide a first output during a forward pass. The first set of variables can change in response to feedback representing differences between desired network outputs and actual network outputs. The second subunit has a plurality of second inputs, and a corresponding second set of variables for operating upon the second inputs to provide a second output. The second set of variables can change in response to differences between desired network outputs for selected network inputs and actual network outputs. The computer provides an activating variable representing the difference between current second output and previous second outputs.

Abstract: A layer of chemically compatible conductive material is applied to ferroelectric semiconductor material having a permanent polarization below its Curie temperature. The materials are heated to a temperature above the Curie temperature of the ferroelectric semiconductor material and allowed to cool. The result is a low resistance contact. The ferroelectric semiconductive material may be a layer on a non ferroelectric semiconductor material with a matching work function or matching dopant levels.

Abstract: A computer operated telephone dialing system has a program that selects a telephone number from a directory of names and corresponding telephone numbers. The program sends an off-hook instruction and a select flag to the computer's printer port, followed by a hexidecimal code sequence representing digits of the telephone number. A dialing circuit coupled to said printer port is activated by the select flag and, in response to the hexidecimal code sequence, provides DTMF dialing signals to telephone lines. The program then sends an on-hook signal, allowing a telephone handset to be used.

Abstract: An interactive system transfers video and audio information from a central facility to terminals by means of addressed television frames. Each audio frame includes information representing a single channel of time compressed audio. The terminals store correctly addressed frames. Video frames are repeatedly replayed to provide series of still-frame images. The audio frame is used to generate a normal speed audio output.

Abstract: A frequency synthesizer has in series a digital synthesizer, mix and divide circuits, and a mix only circuit. Each mix and divide circuit includes a mixer having an IF input, a LO input and a RF output, a filter, and a frequency divider. A segment of a digital word selects a tone from a tone generator and also selects a passband of the filter. The selected tone is coupled to the LO input of the mixer. The frequency divider divides the filtered RF output of the mixer. The mix only circuit is at the output of the last of the series connected mix and divide circuits. The mix only circuit is similar to a mix and divide circuit, but does not have a frequency divider. The digital synthesizer provides a signal to the input of the first of the series connected mix and divide circuits, the frequency of which corresponds to a segment of a digital word. The filter is preferably a voltage tunable bandpass filter.

Abstract: A telecommunication system for synchronizing a digital bit stream sent from a central facility to a terminal on lines of television frames. Each active line starts with a horizontal sync pulse and a color burst. A first data clock at the central facility provides first data clock pulses synchronized with said color burst. A circuit at the central facility provides a flag bit one clock pulse wide delayed by a first constant number of clock pulses from the start of the line's horizontal sync pulse. A second data clock at the terminal provides second data clock pulses synchronized with said color burst. A circuit at said terminal provides a time window delayed by a second constant number of clock pulses from the start of the line's horizontal sync pulse so as to bracket said flag bit. A plurality of second data clock pulses occur during said time window.

Abstract: A printed circuit module retainer has two linear guides which are channeled to accept edges of printed circuit module. A serpentine spring transverses each channel to retain an inserted module.

Abstract: A telecommunication system for transmitting video and plural audio information on a transmission medium includes a facility and an addressable terminal. The facility includes arrangements for providing addressed video frames containing information corresponding to still images and for providing addressed audio frames containing information corresponding to audio bursts. There is also means for providing non-addressed signals containing information corresponding to at least one channel of continuous audio. The addressed video frames, addressed audio frames, and non-addressed signals are coupled to the transmission medium and sent to the terminal. The addressable terminal includes means for detecting said addressed video frames, addressed audio frames, and non-addressed signals on said transmission medium. Circuits select video frames and audio frames in response to their corresponding addresses.

Abstract: A telecommunication system has a facility and an addressable terminal coupled through a transmission medium. The facility includes means for providing addressed video frames containing information corresponding to still images and for providing addressed audio frames containing information corresponding to audio bursts. There is also provided for non-addressed signals containing information corresponding to at least two channels of continuous audio. A selection code is inserted in one or more of the frames. The addressed video and audio frames and non-addressed signals are coupled to the transmission medium to the addressable terminal. The terminal has means for detecting said addressed video frames, addressed audio frames, and non-addressed signals. Circuits select video frames and audio frames in response to their corresponding addresses. One of said continuous audio channels is selected in response to said selection code.

Abstract: A method of soldering leadless component carriers or the like to printed circuit boards uses a solder paste containing copper spheres in addition to particulate solder and flux. The solder is melted and the copper spheres are embedded in the resulting solder matrix, offsetting the carrier from the printed circuit board.

Abstract: A liquid electrolyte solution is added to porous components of an electrochemical cell prior to assembly. The electrolyte solution is frozen and all the components are assembled into a sealed cell. The electrolyte solution is then allowed to melt.

Abstract: A voltage tumble filter has several parallel coupled half wavelength lines. Each open end of a line is coupled to ground through a corresponding varactor. Voltage may be distributed to the varactors through a thin wire attached a quarter wavelength from each varactor.

Abstract: A solder paste contains solder plated copper spheres in addition to particulate solder and flux. The paste can be used to solder leadless component carriers or the like to printed circuit boards., After melting the solder, the copper spheres are embedded in the resulting solder matrix, offsetting the carrier from the printed circuit board.