Texas Instruments Corporation

Giulio Porrovecchio

Giulio Giuseppe Marotta

Giuliano Imondi

Giuseppe Savarese

Luciano Talamonti

Eros Pasero

Integrated circuit with a non-volatile variable resistor which is particularly adapted for use in a neuronic network. The integrated circuit comprises a symmetrically replicated structure including a floating gate MOS transistor (TRP; TRN) and an EEPROM memory cell based upon N-channel MOS transistors and including a read-out MOS transistor (TSP; TSN) and a tunnel-effect charge injection MOS element (TUNP; TUNN), with floating gates. The floating gates of all transistors and of the tunnel-effect element are connected together, with the ends of the resistor (R1P, R2P; R1N, R2N) being taken out from the source and drain regions of the floating gate MOS transistor.

Integrated circuit with non-volatile, variable resistor, for use in

Circuitry and a method are provided for selectively switching a negative voltage (-V.sub.nn) to portions of CMOS integrated circuits, which circuitry comprises a switching/decoding matrix. The switching/decoding matrix comprises a control and decode logic (CDL) which controls signal VPPENABLE to control a positive charge pump (PCP) producing positive voltage (+V.sub.pp) and which further controls signal VNNENABLE to control a negative charge pump (NCP) producing said negative voltage (-V.sub.nn). The switching/decoding matrix further comprises, for each line to be switched, a switching module which comprises a PMOS transistor (PS) having its source connected to said line and its drain connected to receive said negative voltage (-V.sub.nn) produced by said negative charge pump (NCP). The PMOS transistor (PS) gate is driven by a drive circuit being in turn driven by said control and decode logic (CDL) and connected so as to receive the positive voltage (+V.sub.pp) provided by said positive charge pump (PCP).

Circuitry and method for selectively switching negative voltages in CMOS

Data processing system implementing architecture of a neural network which is subject to a learning process, wherein the data processing system includes n.times.n synapses arranged in an array of j rows and i columns. A plurality of operational amplifiers respectively corresponding to the rows of the array are provided, with each operational amplifier defining a neuron. The input terminals of all of the synapses arranged in a respective column of the array are connected together and define n inputs of the neural network. The output terminals of the synapses arranged in a respective row of the array are connected together and serve as the inputs to a corresponding one of the plurality of operational amplifiers. Each synapse includes a capacitor connected between ground potential and the input terminal for weighting the synapse by storing a weighting voltage applied thereto. A random access memory has digitally stored voltage values for weighting all of the synapses. A plurality of digital-analog converters, one for each column of the array of synapses, are connected to the random access memory for converting the digital voltage values for weighting the synapses into analog voltage values. The digital-analog converters provide respective outputs to the weighting terminals of the synapses of a column via respective electronic switches for each synapse. Each row of the array includes a bistable circuit for driving the respective electronic switches under the control of a control section which also provides function commands and data to the random access memory.

Data processing system for implementing architecture of neural network

The subject of the invention is a learning system for a neural net physically insertable in the learning process, which comprises a detecting member for presenting to said neural net the basic information set that said neural net has to learn in order to provide a desired response; a microprocessor suitable to iteratively execute a learning algorithm based on a comparison among said basic information set itself, the response that the neural net provides and the response that one wants to obtain from the neural net (see FIG. 1).

Growing community of inventors

Rieti, Italy

Giulio Porrovecchio