Abstract: A multiplexing circuit for a memory, including: a first parallel branch for coupling a program voltage to a first bitline corresponding to a first bit cell of the memory during a program mode of the memory; and a second parallel branch for coupling a program inhibit voltage to a plurality of additional bitlines corresponding to a plurality of additional bit cells of the memory during a program inhibit mode of the memory, wherein the first parallel branch couples an erase inhibit voltage to the plurality of additional bitlines during an erase inhibit mode of the memory, and wherein the second parallel branch couples an erase voltage to the first bitline during an erase mode of the memory.

Abstract: A computer-implemented method of generating a security language query from a user input query includes receiving, at a computer system, an input security hunting user query indicating a user intention; selecting, using a trained machine learning model and based on the input security hunting query, an example user security hunting query and corresponding example security language query; generating, using the trained machine learning model, query metadata from the input security hunting query; generating a prompt, the prompt comprising: the input security hunting user query; the selected example user security hunting query and the corresponding example security language query; and the generated query metadata; inputting the prompt to a large language model; receiving a security language query from the large language model corresponding to the input security hunting query reflective of the user intention.

Abstract: Techniques are described herein that are capable of detecting an algorithmic attack against a hosted artificial intelligence (AI) system based on inputs (e.g., queries) and outputs of the hosted AI system. In a first example, a feature-based classifier model is used to generate a classification score based on features that are derived from numerical representations of the queries and the outputs, and an algorithmic attack is detected based on the classification score being greater than or equal to a score threshold. In a second example, a transformer-based model is used to generate a vector by providing a multivariate time series, which is based on attribute(s) of the inputs and attribute(s) of the outputs, as an input to the transformer-based model, and an algorithmic attack is detected based on a distance between the vector and a point corresponding to a reference vector being less than or equal to a distance threshold.

Abstract: A data storage device having a flash translation layer configured to handle file-system defragmentation in a manner that substantially avoids physical data movement in a flash memory. In an example embodiment, a memory controller operates to update a logical-to-physical table thereof to change association of physical addresses of sections of user data from being associated with source logical addresses to being associated with destination logical addresses of the host defragmentation requests without moving the user data in the flash memory. Such updates can reduce the number of instances in which the host addresses a non-contiguous logical-address range, which results in a beneficial reduction of the number of input/output commands sent to the data storage device and of the associated processing overhead.

Abstract: A system includes a memory and a processor. The memory stores a set of videos each of which is associated with an application of a set of applications and depicts a process performed using the application. For each video, the processor generates metadata associated with the depicted process, and stores the metadata within a collection of metadata in the memory. The processor receives a request for information from a user. In response, the processor identifies an application of the set of applications for use in generating a response to the request. The processor generates, based on the collection of metadata, instructions for interfacing with the first application to generate the response. The processor executes the instructions to interface with the first application to generate the response, and transmits the response to the user.

Abstract: Systems, computer program products, and methods are described herein for implementing parametric optimization analysis for resource selection. The present invention is configured to determine a first set of requirements associated with a resource exchange agreement; identify one or more non-fungible tokens (NFTs) for one or more categories of past resource exchange agreements based on at least the first set of requirements; extract, from the one or more NFTs, one or more resource descriptors associated with one or more past resource exchange agreements in the one or more categories; predict, using a machine learning subsystem, an optimal resource valuation model for one or more resources that meet the first set of requirements using the one or more resource descriptors and the first set of requirements; and transmit control signals configured to cause a first end-point device to display the optimal resource valuation model.

Abstract: Disclosed are a sense circuit and memory structure incorporating the sense circuit. The sense circuit is connected to voltage rails at VDD1 and VDD2, respectively, where VDD2˜½*VDD1. During a sensing operation, VDD1 provides power to develop a voltage differential between Vdata and Vref on sense nodes. A voltage comparator samples Vdata and Vref and, based on a detectable voltage differential (minVdiff), outputs a data output value. To increase the speed at which minVdiff is reached, an equalization process is performed at the initiation of the sensing operation and includes using pre-charge transistors to quickly equalize the sense nodes to VDD2. Following equalization, Vdata and Vref only need to be pulled up or down from VDD2. Thus, minVdiff is reached faster and sampling by the voltage comparator can be performed earlier in time, reducing the overall time required for performing the sensing operation and for powering the sense circuit.

Abstract: There is provided a system and method for angstrom confinement of trapped ions. The method including: receiving water molecules and ionic compounds in a first reservoir, an angstrom confinement assembly is positioned between the first reservoir and a second reservoir, the angstrom confinement assembly defining angstrom conduits; and repeatedly applying an electric field across a first electrode and a second electrode, the first electrode on a same side of the angstrom confinement assembly as the first reservoir and the second electrode on a same side of the angstrom confinement assembly as the second reservoir, the electric field applied such that, when the electric field is applied, positive ions of the ionic compounds are induced to flow through the angstrom conduits, and wherein, when the electric field is not applied, water molecules flow into the angstrom conduits due to capillary forces to confine the positive ions in the angstrom conduits.

Abstract: Aspects of the disclosure relate to an intelligent resource evaluation engine. A computing platform may monitor the plurality of RPA machines to detect parameter information. The computing platform may store the parameter information along with corresponding RPA machines as a key value pairs in a database. The computing platform may identify first current parameter information for a first RPA machine using the key value pairs. The computing platform may input the first current parameter information into an intelligent resource evaluation model, which may output first machine selection information for the first RPA machine. Based on identifying that the first RPA machine is sufficient to execute the first robotic automation process, the computing platform may send direct the first RPA machine to execute the first robotic automation process.

Abstract: A system includes a machine learning (ML) engine in which a data store is coupled to a processing system. The processing system executes code to receive a dataset from the data store, and produces three models each comprising three tests. Each of the tests seeks to detect one of three anomaly types corresponding to each of the models. The processing system performs at least two of the three tests relating to each of the three anomaly types. Separately for each anomaly type, the processing system detects an anomaly when two-out-of-three (2oo3) tests conclude that the anomaly is present in the dataset. The dataset including the flagged anomalies is stored in a data repository. The anomaly is filtered from the dataset. The processing system is configured to use data from the dataset to retrain an existing trained ML model.

Abstract: Embodiments of the disclosure provide memory circuit, a sense amplifier and associated method for reading a resistive state in a memory device. The sense amplifier includes a bit cell configurable to a high or low resistance state; a sensing circuit that detects a voltage drop across the bit cell in response to an applied read current during a read operation and generates a high or low logic output at an output node; and a pulse generation circuit that increases the applied read current with an injected current pulse when a low to high transition of the resistive state of the bit cell is detected.

Abstract: A system for generating an expandable pool of virtual agents to provide solutions in a virtual environment comprises a call center server in communication with a metaverse server. The call center server generates the virtual agents with corresponding skills of each real-world agent using a machine learning model based on historical observation operations performed by each particular real-world agent and a historical rule set to address the service issue. The metaverse server identifies the virtual agents in the virtual environment. The call center server routes a service inquiry with a service issue from a user equipment of a user to the metaverse server. The metaverse server identifies a first virtual agent with a first skill and conducts a first interaction between the first virtual agent and a user avatar to generate the customized solution to address the service issue.

Abstract: Described herein is a graphics processor having processing resources with configurable thread and register configurations. Program code can configure a number of registers and accumulators that will be used by hardware threads during execution of the program code by the graphics processor. Processing resources within the graphics processor can be configured to assign different numbers of registers and accumulators to hardware threads based on the configuration requested by program code to be executed by the processing resource.

Abstract: A VCSEL array includes adjacent first and second VCSEL elements and a trench surrounding the VCSEL arrays. The first and second VCSEL elements include a first output window and a second output window. The first VCSEL element includes a part of a metal layer shared by the first and second VCSEL elements. The distance between the inner edge of the trench and the center of the first output window is larger than half the distance between centers of the first and second output windows.

Abstract: A system for generating an expandable pool of virtual agents to provide solutions in a virtual environment comprises a call center server in communication with a metaverse server. The call center server generates the virtual agents with corresponding skills of each real-world agent using a machine learning model based on historical observation operations performed by each particular real-world agent and a historical rule set to address the service issue. The metaverse server identifies the virtual agents in the virtual environment. The call center server routes a service inquiry with a service issue from a user equipment of a user to the metaverse server when determining that a particular real-world agent is not available. The metaverse server identifies a first virtual agent with a first skill and conducts a first interaction between the first virtual agent and a user avatar to generate the customized solution to address the service issue.

Abstract: A structure includes a level shifter, first and second variable voltage generators, and a programmable voltage generator. The level shifter includes low and high supply voltage nodes and two parallel branches, including multiple transistors, connected between the nodes. The programmable voltage generator generates and applies a high supply voltage (VH) to the high supply voltage node. VH is programmable to one of multiple possible VH levels. Based on the voltage level of VH, the first variable voltage generator generates and applies a low supply voltage (VL) to the low supply voltage node and the second variable voltage generator generates and applies a gate bias voltage (VGB) to gates of some transistors. By tracking VH and adjusting VL and VGB based on thereon, the voltage level shifter operates within the SOA at high VHs, remains operable at low VHs, and maintains operating speed at mid-level VHs.

Abstract: Aspects of the disclosure relate to an intelligent resource evaluation engine. A computing platform may monitor the plurality of RPA machines to detect parameter information. The computing platform may store the parameter information along with corresponding RPA machines as a key value pairs in a database. The computing platform may identify first current parameter information for a first RPA machine using the key value pairs. The computing platform may input the first current parameter information into an intelligent resource evaluation model, which may output first machine selection information for the first RPA machine. Based on identifying that the first RPA machine is sufficient to execute the first robotic automation process, the computing platform may send direct the first RPA machine to execute the first robotic automation process.

Abstract: Methods and Wi-Fi devices that utilize the target wait time (TWT) feature of the Wi-Fi specification are disclosed. These 5 methods reduce power consumption by selectively listening to only a portion of the beacons that are transmitted by the access point. In one scenario, the Wi-Fi device only listens to one beacon per TWT wake interval. In another embodiment, the Wi-Fi device may have knowledge of the application that is being executed, such as its allowable latency. The Wi-Fi device may use this allowable latency to determine which when to exit low power mode to receive a beacon. Further, mechanisms to ensure that the connection between the Wi-Fi device and the access point are also disclosed.

Abstract: Systems, computer program products, and methods are described herein for intelligent workflow management in robotic process automation (RPA). The present disclosure comprises a workflow replication subsystem configured to replicate a sequence of actions executed by an RPA bot in an application server; a fog computing subsystem operatively coupled to the workflow replication subsystem, wherein the fog computing subsystem is configured to extract metadata associated with execution of the sequence of actions; and an anomaly resolution subsystem operatively coupled to the fog computing subsystem and the workflow replication subsystem, wherein the anomaly resolution subsystem is configured to troubleshoot instances of interruptions in the execution of the sequence of actions, wherein the instances of interruptions comprise at least potential failures in at least one action yet to be executed by the RPA bot.

Abstract: Systems, computer program products, and methods are described herein for securing and emulating computing system configurations.