The present disclosure includes three dimensional memory arrays, and methods of processing the same. A number of embodiments include a plurality of conductive lines separated from one other by an insulation material, a plurality of conductive extensions arranged to extend substantially perpendicular to the plurality of conductive lines, and a storage element material formed around each respective one of the plurality of conductive extensions and having two different contacts with each respective one of the plurality of conductive lines, wherein the two different contacts with each respective one of the plurality of conductive lines are at two different ends of that respective conductive line.

Apparatuses and methods can be related to implementing a binary neural network in memory. A binary neural network can be implemented utilizing a resistive memory array. The memory array can comprise programmable memory cells that can be programed and used to store weights of the binary neural network and perform operations consistent with the binary neural network. The weights of the binary neural network can correspond to non-zero values.

Disclosed are a neuromorphic device and a unit synapse devices forming the same. The unit synapse device has a learning device and an inference device. The learning device and the inference device may share a via oxide layer and a common electrode, and a learning operation and an inference operation may be performed in one unit synapse device.

In an example, an apparatus can include an array of variable resistance memory cells and a neural memory controller coupled to the array of variable resistance memory cells and configured to apply a sub-threshold voltage pulse to a variable resistance memory cell of the array to change a threshold voltage of the variable resistance memory cell in an analog fashion from a voltage associated with a reset state to effectuate a first synaptic weight change; and apply additional sub-threshold voltage pulses to the variable resistance memory cell to effectuate each subsequent synaptic weight change.

A re-writeable non-volatile memory device including a re-writeable non-volatile two-terminal memory element (ME) having tantalum. The ME including a first terminal, a second terminal, a first layer of a conductive metal oxide (CMO), and a second layer in direct contact with the first layer. The second layer and the first layer being operative to store at least one-bit of data as a plurality of resistive states, and the first and second layer are electrically in series with each other and with the first and second terminals.

A vertical resistive memory array is presented. The array includes a pillar electrode and a switching liner around the side perimeter of the pillar electrode. The array includes two or more vertically stacked single cell (SC) electrodes connected to a first side of the switching liner. The juxtaposition of the switching liner, the pillar electrode, and each SC electrode forms respective resistance switching cells (e.g., OxRRAM cell). A vertical group or bank of these cells may be connected in parallel and each share the same pillar electrode. The cells in the vertical cell bank may written to or read from as a group to limit the effects of inconsistent CF formation of any one or more individual cells within the group.

Methods and apparatuses for thin film transistors and related fabrication techniques are described. The thin film transistors may access two or more decks of memory cells disposed in a cross-point architecture. The fabrication techniques may use one or more patterns of vias formed at a top layer of a composite stack, which may facilitate building the thin film transistors within the composite stack while using a reduced number of processing steps. Different configurations of the thin film transistors may be built using the fabrication techniques by utilizing different groups of the vias. Further, circuits and components of a memory device (e.g., decoder circuitry, interconnects between aspects of one or more memory arrays) may be constructed using the thin film transistors as described herein along with related via-based fabrication techniques.

A computing device for a multidimensional vector neural network includes: input lines to which multidimensional input vectors are input; output lines intersecting the input lines; memory cells disposed at intersecting points between the input lines and the output lines and configured to store weight elements included in multidimensional weight vectors; selectors configured to transmit a value output from each of the output lines to any one of adders; and the adders configured to accumulate values received from the selectors in a predetermined number of cycles, wherein, for each of the multidimensional weight vectors, weight elements included in the multidimensional weight vector are stored in reference memory cells that connect a corresponding single reference input line and corresponding two or more reference output lines.

A resistive element array circuit includes word lines, bit lines, resistive elements, a selector, a differential amplifier, and a ground terminal. The word lines are coupled to a power supply. The resistive elements are each disposed at an intersection of corresponding one of the word lines and corresponding one of the bit lines. The selector is configured to select one word line and one bit line. The differential amplifier includes a positive input terminal configured to be coupled to the selected one of the bit lines which is selected by the selector, a negative input terminal configured to be coupled to non-selected one of the bit lines which is not selected by the selector and to non-selected one of the word lines which is not selected by the selector, an output terminal being coupled to the negative input terminal. The ground terminal is coupled to the positive input terminal.

Methods, systems, and devices for techniques for forming memory structures are described. Forming a memory structure may include etching a stack of material including a conductive line, a first electrode and a sacrificial material to divide the stack of material into multiple sections. The process may further include depositing an oxide material in each of the first quantity of channels to form multiple oxide materials. The sacrificial material may be etched to form a second channel between two oxide materials of the multiple oxide materials. Memory material may be deposited over the two oxide materials and the second channel, which may create a void in the second channel between the memory material and the first electrode. The memory material may be heated to fill the void in the second channel.