An error compensation circuit for analog capacitor memory circuits includes a first transistor and a second transistor with gates connected respectively to top and bottom of an analog memory capacitor to read a voltage charged in the analog memory capacitor; a first switch and a second switch connected respectively to the first transistor and the second transistor to select the voltage to read; a first capacitor and a second capacitor to charge an electric charge to compensate or refresh the analog memory capacitor according to on/off of the first switch and the second switch; and an input terminal connected to sources of the first transistor and the second transistor to apply the voltage to operate the circuit. Accordingly, it is possible to compensate for an unintended phenomenon of the analog capacitor memory or refresh a change in memory value caused by leakage.

In some embodiments, a superconducting parametric amplification neural network (SPANN) includes neurons that operate in the analog domain, and a fanout network coupling the neurons that operates in the digital domain. Each neuron is provided one or more input currents having a resolution of several bits. The neuron weights the currents, sums the weighted currents with an optional bias or threshold current, then applies a nonlinear activation function to the result. The nonlinear function is implemented using a quantum flux parametron (QFP), thereby simultaneously amplifying and digitizing the output current signal. The digitized output of some or all neurons in each layer is provided to the next layer using a fanout network that operates to preserve the digital information held in the current.

There is provided a magnetoresistance effect element includes: a channel layer that extends in a first direction; a recording layer which includes a film formed from a ferromagnetic material, of which a magnetization state is changed to one of two or greater magnetization states, and which is formed on the channel layer; a non-magnetic layer that is provided on a surface of the recording layer; a reference layer which is provided on a surface of the non-magnetic layer, which includes a film formed from a ferromagnetic material, and of which a magnetization direction is fixed; a terminal pair that includes a first terminal and a second terminal which are electrically connected to the channel layer with an interval in the first direction, and to which a current pulse for bringing the recording layer to any one magnetization state with a plurality of pulses is input by flowing a current to the channel layer between the first terminal and the second terminal; and a third terminal that is electrically connected to the reference layer.

Systems, apparatuses, and methods related to comparison of biometric identifiers in memory are described. An example apparatus includes an array of memory cells, a plurality of logic blocks in complementary metal-oxide-semiconductor (CMOS) under the array, and a controller coupled to the array of memory cells. The controller is configured to control a first portion of the plurality of logic blocks to receive a first subset of a set of biometric identifiers from the array and to perform a first comparison operation thereon and control a second portion of the logic blocks to receive a second subset of the set of biometric identifiers from the array and to perform a second comparison operation thereon. The first and second subsets of the biometric identifiers are different biometric identifiers and the first and second comparison operations are performed to determine a match of the first and second subsets respectively to a stored template.

Methods and systems of implementing a convolutional neural network are described. In an example, a structure may receive input signals and distribute the input signals to a plurality of unit cells. The structure may include a plurality of multi-kernel modules that may include a respective set of unit cells. A unit cell may correspond to an element of a kernel being implemented in the convolutional neural network and may include a storage component configured to store a weight of a corresponding element of the kernel. A first pass gate of the unit cell may be activated to pass a stored weight of the unit cell to a plurality of operation circuits in the corresponding unit cell, such that the stored weight may be applied to the input signals. The structure may generate a set of outputs based on the application of the stored weights to the input signals.

Novel connection between neurons of a neural network is provided. A perceptron included in the neural network includes a plurality of neurons; the neuron includes a synapse circuit and an activation function circuit; and the synapse circuit includes a plurality of memory cells. A bit line selected by address information for selecting a memory cell is shared by a plurality of perceptrons. The memory cell is supplied with a weight coefficient of an analog signal, and the synapse circuit is supplied with an input signal. The memory cell multiplies the input signal by the weight coefficient and converts the multiplied result into a first current. The synapse circuit generates a second current by adding a plurality of first currents and converts the second current into a first potential. The activation function circuit is a semiconductor device that converts the first potential into a second potential by a ramp function and supplies the second potential as an input signal of the synapse circuit included in the perceptron in a next stage.

Technology for reconfigurable input precision in-memory computing is disclosed herein. Reconfigurable input precision allows the bit resolution of input data to be changed to meet the requirements of in-memory computing operations. Voltage sources (that may include DACs) provide voltages that represent input data to memory cell nodes. The resolution of the voltage sources may be reconfigured to change the precision of the input data. In one parallel mode, the number of DACs in a DAC node is used to configure the resolution. In one serial mode, the number of cycles over which a DAC provides voltages is used to configure the resolution. The memory system may include relatively low resolution voltage sources, which avoids the need to have complex high resolution voltage sources (e.g., high resolution DACs). Lower resolution voltage sources can take up less area and/or use less power than higher resolution voltage sources.

A method for implementing a neural network system in an integrated circuit includes presenting digital pulses to word line inputs of a matrix vector multiplier including a plurality of word lines, the word lines forming intersections with a plurality of summing bit lines, a programmable Vt transistor at each intersection having a gate connected to the intersecting word line, a source connected to a fixed potential and a drain connected to the intersecting summing bit line, each digital pulse having a pulse width proportional to an analog quantity. During a charge collection time frame charge collected on each of the summing bit lines from current flowing in the programmable Vt transistor is summed. During a pulse generating time frame digital pulses are generated having pulse widths proportional to the amount of charge that was collected on each summing bit line during the charge collection time frame.

A near memory system is provided for the calculation of a layer in a machine learning application. The near memory system includes an array of memory cells for storing an array of filter weights. A multiply-and-accumulate circuit couples to columns of the array to form the calculation of the layer.

A machine learning system and method. The machine learning system includes at least one computation circuit that performs a weighted summation of incoming signals and provides a resulting signal. The weighted summation is carried out at least in part by a magnetic element in which weights are adjusted based on changes in effective magnetic susceptibility of the magnetic element.