Abstract: Automatic hand gesture determination may be a challenging task considering the complex anatomy and high dimensionality of the human hand. Disclosed herein are systems, methods, and instrumentalities associated with recognizing a hand gesture in spite of the challenges. An apparatus in accordance with embodiments of the present disclosure may use machine learning based techniques to identify the area of an image that may contain a hand and to determine an orientation of the hand relative to a pre-defined direction. The apparatus may then adjust the area of the image containing the hand to align the orientation of the hand with the pre-defined direction and/or to scale the image area to a pre-defined size. Based on the adjusted image area, the apparatus may detect a plurality of hand landmarks and predict a gesture indicated by the hand based on the plurality of detected landmarks.

Abstract: Transistor structures with a channel semiconductor material that is passivated with two-dimensional (2D) crystalline material. The 2D material may comprise a semiconductor having a bandgap offset from a band of the channel semiconductor. The 2D material may be a thin as a few monolayers and have good temperature stability. The 2D material may be a conversion product of a sacrificial precursor material, or of a portion of the channel semiconductor material. The 2D material may comprise one or more metal and a chalcogen. The channel material may be a metal oxide semiconductor suitable for low temperature processing (e.g., IGZO), and the 2D material may also be compatible with low temperature processing (e.g., <450° C.). The 2D material may be a chalcogenide of a metal present in the channel material (e.g., ZnSx or ZnSex) or of a metal absent from the channel material when formed from a sacrificial precursor.

Abstract: The present disclosure provides a system and a method for controlling motion of a bank of elevators. The method includes accepting current requests for service by the bank of elevators, accepting a partial trajectory of a motion of a person moving in an environment serviced by the bank of elevators, and obtaining a probability of a future elevator request. The method further includes processing the partial trajectory with a neural network trained to estimate a weighted combination of probability density functions that indicates an arrival time distribution of the person, and generating a set of possible future requests jointly representing the probability of the future elevator request and the arrival time distribution. The method further includes optimizing a schedule of the bank of elevators to serve the current requests and the set of possible future requests, and controlling the bank of elevators according to the schedule.

Abstract: Systems, computer program products, and methods are described herein for auto-scaling volatile memory allocation in an electronic network. The present invention is configured to access metadata of at least one volatile memory component, wherein the metadata is associated with at least one application; determine a current volatile memory allocation for the metadata; determine a current metadata format of the metadata; apply the metadata to a volatile memory allocation machine learning model; generate, based on the application of the metadata to the volatile memory allocation machine learning model, a new volatile memory allocation for the metadata; and apply the new volatile memory allocation to the metadata of the at least one volatile memory component, wherein the application of the new volatile memory allocation to the metadata comprises at least one of an upscaling, a downscaling, or a constant.

Abstract: An efficient state-machine-based pattern matching technique processes tokens in an input queue and identifies patterns in the sequence of tokens that match one or more predetermined input patterns without backtracking. Tokens can include data or no data and a time component. The tokens can be a stream of data generated by a sensor, which transforms a physical property into a digital quantity. The pattern matching technique processes the input queue in a single direction, and does not examine any previously examined token. In an implementation, specific patterns to be matched are specified using a state machine, where the state machine is specified in a state table and operates using a state stack.

Abstract: An edge computing platform with machine learning capability is provided between a local network with a plurality of sensors and a remote network. A machine learning model is created and trained in the remote network using aggregated sensor data and deployed to the edge platform. Before being deployed, the model is edge-converted (“edge-ified”) to run optimally with the constrained resources of the edge device and with the same or better level of accuracy. The “edge-ified” model is adapted to operate on continuous streams of sensor data in real-time and produce inferences. The inferences can be used to determine actions to take in the local network without communication to the remote network. A closed-loop arrangement between the edge platform and remote network provides for periodically evaluating and iteratively updating the edge-based model.

Abstract: Systems, computer program products, and methods are described herein for auto-scaling volatile memory allocation in an electronic network. The present invention is configured to access metadata of at least one volatile memory component, wherein the metadata is associated with at least one application; determine a current volatile memory allocation for the metadata; determine a current metadata format of the metadata; apply the metadata to a volatile memory allocation machine learning model; generate, based on the application of the metadata to the volatile memory allocation machine learning model, a new volatile memory allocation for the metadata; and apply the new volatile memory allocation to the metadata of the at least one volatile memory component, wherein the application of the new volatile memory allocation to the metadata comprises at least one of an upscaling, a downscaling, or a constant.

Abstract: A computer includes a processor and a memory storing instructions executable by the processor to receive sensor data indicating a current position of an object, determine a predicted position of the object at a future time, and instruct a component of a vehicle to actuate based on the current position being in a first zone of a plurality of zones surrounding the vehicle and the predicted position being in a second zone of the plurality of zones different than the first zone. The zones are nonoverlapping and have preset boundaries relative to the vehicle.

Abstract: A system for providing performance optimization for a software solution may scan multiple predefined levels of the software solution to extract corresponding metadata information from each of the multiple predefined levels. The system may store the extracted corresponding metadata information pertaining to standard parameters associated with performance of the software solution. The system may determine a standard score based on a plurality of attributes of the extracted corresponding metadata information, optimize the determined standard score based on training data received from a learning model, and generate an insight information comprising information related to determined rule violations and of evaluation steps involved in determining the determined standard score.

Abstract: The shape and/or location of an organ may change in accordance with changes in the body shape and/or pose of a patient. Described herein are systems, methods, and instrumentalities for automatically determining, using an artificial neural network (ANN), the shape and/or location of the organ based on human models that reflect the body shape and/or pose the patient. The ANN may be trained to learn the spatial relationship between the organ and the body shape or pose of the patient. Then, at an inference time, the ANN may be used to determine the relationship based on a first patient model and a first representation (e.g., a point cloud) of the organ so that given a second patient model thereafter, the ANN may automatically determine the shape and/or location of the organ corresponding to the body shape or pose of the patient indicated by the second patient model.

Abstract: Data samples are transmitted from a central server to at least one local server apparatus. The central server receives a set of predictions from the at least one local server apparatus that are based on the transmitted set of data samples. The central server trains a central model based on the received set of predictions. The central model, or a portion of the central model corresponding to a task of interest, can then be sent to the at least one local server apparatus. Neither local data from local sites nor trained models from the local sites are transmitted to the central server. This ensures protection and security of data at the local sites.

Abstract: A method includes processing event data to detect a status of a network function. The event data is processed based on two or more conditions defined by a correlation policy. The correlation policy includes a non-deterministic finite automata tree (NFAT) structure correlation policy having a policy type. The method also includes determining a first value of a first condition of the two or more conditions. The method further includes determining a second value of a second condition of the two or more conditions. The method additionally includes determining the policy type of the NFAT structure correlation policy. The method also includes determining whether the first value is greater than a first preset value indicative of whether the first condition is satisfied. The method further includes determining whether the second value is greater than a second preset value indicative of whether the second condition is satisfied.

Abstract: Automatically validating the calibration of an visual sensor network includes acquiring image data from visual sensors that have partially overlapping fields of view, extracting a representation of an environment in which the visual sensors are disposed, calculating one or more geometric relationships between the visual sensors, comparing the calculated one or more geometric relationships with previously obtained calibration information of the visual sensors, and verifying a current calibration of the visual sensors based on the comparison.

Abstract: Systems and methods are provided for monitoring and/or controlling operations of an artificial lift system of a production well, which employ a gateway device disposed at a wellsite corresponding to the well, and at least one remote cloud-computing system operably coupled to the gateway device. The gateway device includes at least one first interface to the artificial lift system, at least one second interface to the at least one cloud-computing system, and a processor configured to execute a first application that acquires data produced by the artificial lift system and communicated to the gateway device via the first interface.

Abstract: A system includes processing circuitry and a memory connected to the processing circuitry. The memory is configured to store executable instructions that, when executed by the processing circuitry, facilitate performance of operations. The operations include filtering, based upon business logic, business data within a cache database. Converting the filtered business data into a data model. Loading the data model to a persistent cache database. Obtaining event messages from a data source, where the event messages are generated by one or more state changes within a network operatively connected to the system. Obtaining, for an event message, event-related data from the data model, where the event-related data is topologically related to the event message. Combining the event message with the event-related data from the data model. Framing the event message with the event-related data; and route the frame according to a user-defined configuration file.

Abstract: In some implementations, a device may receive extensibility data related to one or more custom code objects installed in a current environment. The device may classify the one or more custom code objects in one or more respective categories and determine one or more respective complexities associated with the one or more custom code objects based on the extensibility data. The device may generate an extensibility recommendation for deploying the one or more custom code objects to a target environment based on the one or more respective categories and the one or more respective complexities associated with the one or more custom code objects. The extensibility recommendation may be generated based on the one or more custom code objects satisfying extensibility conditions associated with the target environment. The device may provide an output relating to the extensibility recommendation.

Abstract: A method that includes receiving an input at an interactive conversation service that uses an intent classification model. The method may further include generating, using an encoder model of the intent classification model, a set of output vectors corresponding to the input, where the encoder model is configured to determine a set of metrics corresponding to intent classifications. The method may further include determining, using an outlier detection model of the intent classification model, whether the input is in-domain or out-of-domain (OOD) based on a first vector of the set of output vectors satisfying a domain threshold relative to one or more of the intent classifications. The method may further include outputting, by the intent classification model, a second vector of the set of output vectors that indicates the set of metrics corresponding to the intent classifications or an indication that the input is OOD.

Abstract: A system includes a computer including a processor and a memory. The computer is programmed to determine a target location for a vehicle; based on a vehicle speed and a distance to the target location, determine a tunable acceleration parameter; determine a constraint, based on the tunable acceleration parameter, for a control barrier function to that outputs an acceleration to stop the vehicle at the target location; and upon solving the control barrier function to satisfy the constraint, actuate the vehicle to decelerate based on the acceleration output from the control barrier function.

Abstract: An edge computing platform with machine learning capability is provided between a local network with a plurality of sensors and a remote network. A machine learning model is created and trained in the remote network using aggregated sensor data and deployed to the edge platform. Before being deployed, the model is edge-converted (“edge-ified”) to run optimally with the constrained resources of the edge device and with the same or better level of accuracy. The “edge-ified” model is adapted to operate on continuous streams of sensor data in real-time and produce inferences. The inferences can be used to determine actions to take in the local network without communication to the remote network. A closed-loop arrangement between the edge platform and remote network provides for periodically evaluating and iteratively updating the edge-based model.

Abstract: A system performs named entity recognition for performing natural language processing, for example, for conversation engines. The system uses context information in named entity recognition. The system includes the context of a sentence during model training and execution. The system generates high quality contextual data for training NER models. The system utilizes labeled and unlabeled contextual data for training NER models. The system provides NER models for execution in production environments. The system uses heuristics to determine whether to use a context-based NER model or a simple NER model that does not use context information. This allows the system to use simple NER models when the likelihood of improving the accuracy of prediction based on context is low.