Abstract: A neural network is implemented as a memristive neuromorphic circuit that includes a neuron circuit and a memristive device connected to the neuron circuit. An input voltage is sensed at a first terminal of a memristive device during a feedforward operation of the neural network. An error voltage is sensed at a second terminal of the memristive device during an error backpropagation operation of the neural network. In accordance with a training rule, a desired conductance change for the memristive device is computed based on the sensed input voltage and the sensed error voltage. Then a training voltage is applied to the memristive device. Here, the training voltage is proportional to a logarithmic value of the desired conductance change.

Abstract: A neural network is implemented as a memristive neuromorphic circuit that includes a neuron circuit and a memristive device connected to the neuron circuit. A conductance balanced voltage pair is provided for the memristive device, where the conductance balanced voltage pair includes a set voltage for increasing the conductance of the memristive device and a reset voltage for decreasing the conductance of the memristive device. Either the set voltage and reset voltage, when applied to the memristive device, effects a substantially same magnitude conductance change in the memristive device over a predetermined range of conductance of the memristive device. The provided voltage pair is stored as a conductance balanced map. A training voltage based on the conductance balanced map is applied to the memristive device to train the neural network.

Abstract: A neural network is implemented as a memristive neuromorphic circuit that includes a neuron circuit and a memristive device connected to the neuron circuit. An input voltage is sensed at a first terminal of a memristive device during a feedforward operation of the neural network. An error voltage is sensed at a second terminal of the memristive device during an error backpropagation operation of the neural network. In accordance with a training rule, a desired conductance change for the memristive device is computed based on the sensed input voltage and the sensed error voltage. Then a training voltage is applied to the memristive device. Here, the training voltage is proportional to a logarithmic value of the desired conductance change.

Abstract: A neural network is implemented as a memristive neuromorphic circuit that includes a neuron circuit and a memristive device connected to the neuron circuit. A conductance balanced voltage pair is provided for the memristive device, where the conductance balanced voltage pair includes a set voltage for increasing the conductance of the memristive device and a reset voltage for decreasing the conductance of the memristive device. Either the set voltage and reset voltage, when applied to the memristive device, effects a substantially same magnitude conductance change in the memristive device over a predetermined range of conductance of the memristive device. The provided voltage pair is stored as a conductance balanced map. A training voltage based on the conductance balanced map is applied to the memristive device to train the neural network.

Abstract: An ionic device includes a layer of an ionic conductor containing first and second species of impurities. The first species of impurity in the layer is mobile in the ionic conductor, and a concentration profile of the first species determines a functional characteristic of the device. The second species of impurity in the layer interacts with the first species within the layer to create a structure that limits mobility of the first species in the layer.

Abstract: A device contains a first layer, a second layer; and a membrane between the first and second layers. Mobile ions are in at least one of the first and second layers, and the membrane is permeable to the ions. Interfaces of the conductive membrane with the first layer and the second layer are such that charge of a polarity of the ions collects at the interfaces.

Abstract: A memory device (100) includes a semiconductor wire including a source region (132), a drain region (134), and a channel region (130) between the source region (132) and the drain region (134). A gate structure that overlies the channel region includes a memristive portion (120) and a conductive portion (110) overlying the memristive portion (120).

Abstract: An ionic device includes a layer (220) of an ionic conductor containing first and second species (222, 224) of impurities. The first species (222) of impurity in the layer (220) is mobile in the ionic conductor, and a concentration profile of the first species (222) determines a functional characteristic of the device (200). The second species (224) of impurity in the layer (220) interacts with the first species (222) within the layer (220) to create a structure (226) that limits mobility of the first species (222) in the layer (220).

Abstract: A device contains a first layer (420), a second layer (440); and a membrane (430) between the first and second layers (420, 440). Mobile ions (425) are in at least one of the first and second layers (420, 440), and the membrane (430) is permeable to the ions. Interfaces of the conductive membrane (430) with the first layer (420) and the second layer (440) are such that charge of a polarity of the ions (425) collects at the interfaces.

Abstract: A memory device (100) includes a semiconductor wire including a source region (132), a drain region (134), and a channel region (130) between the source region (132) and the drain region (134). A gate structure that overlies the channel region includes a memristive portion (120) and a conductive portion (110) overlying the memristive portion (120).

Abstract: A memristor crossbar array and method of making employ an interstitial insulator. The memristor crossbar array includes a plurality of memristors in an array. The memristors include columns of memristor material disposed between and connecting to a first plurality of wire electrodes and a second plurality of wire electrodes at cross points between the respective wire electrodes. The memristor crossbar array further includes an insulator of a solid material in an interstitial space between the wire electrodes of the first plurality and between the columns of memristor material. The insulator isolates the memristors from one another and has a dielectric constant that is lower than a dielectric constant of the memristor material. The method of making includes forming the plurality of memristors and filling the interstitial space between adjacent memristors with the insulator material.