Various embodiments of the present disclosure are directed towards a memory cell comprising a high electron affinity dielectric layer at a bottom electrode. The high electron affinity dielectric layer is one of multiple different dielectric layers vertically stacked between the bottom electrode and a top electrode overlying the bottom electrode. Further, the high electrode electron affinity dielectric layer has a highest electron affinity amongst the multiple different dielectric layers and is closest to the bottom electrode. The different dielectric layers are different in terms of material systems and/or material compositions. It has been appreciated that by arranging the high electron affinity dielectric layer closest to the bottom electrode, the likelihood of the memory cell becoming stuck during cycling is reduced at least when the memory cell is RRAM. Hence, the likelihood of a hard reset/failure bit is reduced.

Some embodiments include a memory device and methods of forming the memory device. One such memory device includes a first group of memory cells, each of the memory cells of the first group being formed in a cavity of a first control gate located in one device level of the memory device. The memory device also includes a second group of memory cells, each of the memory cells of the second group being formed in a cavity of a second control gate located in another device level of the memory device. Additional apparatus and methods are described.

Methods, systems, and devices for tapered memory cell profiles are described. A tapered profile memory cell may mitigate shorts in adjacent word lines, which may be leveraged for accurately reading a stored value of the memory cell. The memory device may include a self-selecting memory component with a bottom surface and a top surface opposite the bottom surface. In some cases, the self-selecting memory component may taper from the bottom surface to the top surface. In other examples, the self-selecting memory component may taper from the top surface to the bottom surface. The top surface of the self-selecting memory component may be coupled to a top electrode, and the bottom surface of the self-selecting memory component may be coupled to a bottom electrode.

An operation method for a memory device is provided. The memory device includes a two-terminal selector and a resistance variable storage element coupled to the two-terminal selector. The method includes providing a voltage pulse to the memory device. A voltage applied across the two-terminal selector during a falling part of the voltage pulse falls below a holding voltage of the two-terminal selector. A voltage falling rate of the falling part at which the voltage applied across the two-terminal selector reaches the holding voltage is raised for reducing threshold voltage drift of the two-terminal selector.

Provided herein may be a resistive memory device and a method of operating the resistive memory device. The resistive memory device may include strings coupled between one or more source lines and one or more bit lines, each string including a set of one or more resistive memory cells, one or more word lines respectively coupled to the set of one or more resistive memory cells; and a voltage generator configured to control a level of a turn-on voltage to be applied to one or more unselected word lines among the one or more word lines depending on a program target state of a subset of resistive memory cells including one or more resistive memory cells selected from among the set of one or more resistive memory cells.

An in-memory computing architecture is disclosed that can evaluate the transitive closure of graphs using the natural parallel flow of information in 3-D nanoscale crossbars. The architecture can be implemented using 3-D crossbar architectures with as few as two layers of 1-diode 1-resistor (1D1R) interconnects. The architecture avoids memory-processor bottlenecks and can hence scale to large graphs. The approach leads to a runtime complexity of O(n.sup.2) using O(n.sup.2) memristor devices. This compares favorably to conventional algorithms with a time complexity of O((n.sup.3)/p+(n.sup.2) log p) on p processors. The approach takes advantage of the dynamics of 3-D crossbars not available on 2-D crossbars.

Crossbar arrays with reduced disturbance and methods for programming the same are disclosed. In some implementations, an apparatus comprises: a plurality of rows; a plurality of first columns; a plurality of second columns; a plurality of devices. Each of the plurality of devices is connected among one of the plurality of rows, one of the plurality of first columns, and one of the plurality of second columns. The device further comprises a shared end on the plurality of first columns or the plurality of the second columns connecting to the plurality of the devices in the same row or column; the shared end is grounding or holds a stable voltage potential. In some implementations, one of the devices is: a RRAM, a floating date, a phase change device, an SRAM, a memristor, or a device with tunable resistance. In some implementations the stable voltage potential is a constant DC voltage.

An output, representing synaptic weights of a neural network can be received from first memory elements. The output can be compared to a known correct output. A random number can be generated with a tuned bias via second memory elements. The weights can be updated based on the random number and a difference between the output and the known correct output.

According to one embodiment, in a nonvolatile semiconductor memory device, in a cell block, a local bit line is connected to a bit line via a select transistor. The local bit line extends in a third direction. A local source line is connected to a source line and extends in the third direction. A plurality of memory cells are connected in parallel between the local source line and the local bit line. Each of the memory cells includes a cell transistor and a resistance change element. The cell transistor has a gate connected to a corresponding one of the word lines and one end connected to one of the local bit line or the local source line. The resistance change element is connected between the other end of the cell transistor and the other one of the local bit line or the local source line.

A memory device includes first to nth decks respectively coupled to first to nth row lines which are stacked over a substrate in a vertical direction perpendicular to a surface of the substrate, n being a positive integer, a first connection structure extending from the substrate in the vertical direction to be coupled to the first row line, even-numbered connection structures extending from the substrate in the vertical direction and respectively coupled to ends of even-numbered row lines among the second to nth row lines, and odd-numbered connection structures extending from the substrate in the vertical direction and respectively coupled to ends of odd-numbered row lines among the second to nth row lines. The even-numbered connection structures are spaced apart from the odd-numbered connection structures with the first row line and the first connection structure that are interposed between the even-numbered connection structures and the odd-numbered connection structures.