Abstract: Systems, methods, and computer program products are provided for generating and improving multitask learning models. An example method includes determining a first accuracy metric based on at least two machine learning models performing a plurality of tasks, receiving a multitask learning model including at least one shared layer and a plurality of task-specific layers, determining a second accuracy metric based on the multitask learning model having a first number of shared layers, determining a third accuracy metric based on the multitask learning model having a second number of shared layers, comparing the accuracy metrics, repeating until at least one termination condition is satisfied, and determining a target number of shared layers for the multitask learning model based on at least one of the second accuracy metric, the third accuracy metric, the first number of shared layers, the second number of shared layers, or any combination thereof.

Abstract: A device for indicating the passage of a pig beyond a predetermined location in a pipeline comprising a pig detector connected to a pipeline and communicating with the interior of the pipeline through an opening at the location. The device may comprise a trigger mounted in the pig detector and adapted to project into the interior of the pipeline, an indicator housing located at a distance from the pig detector; an indicator being vertically slidably mounted within the indicator housing; and a push/pull control cable with a first end attached to the trigger and a second end attached to the indicator, such that, when a pig moves in the pipeline past the trigger and actuates the trigger, the push/pull control cable allows the indictor to move upwardly beyond the indicator housing so as to provide a visual indication that the pig has passed by.

Abstract: A nicotine e-vaping device includes a heater, a power control circuit, and a memory module. The heater element is configured to heat nicotine pre-vapor formulation. The power control circuit is coupled to the heater element through a wire. The power control circuit is configured to apply a pulse width modulated power signal to the heater element through the wire, and to receive information over the wire. The memory module is configured to detect a plurality of pulses in the pulse width modulated power signal, record information based on the detected plurality of pulses, and output the recorded information to the power control circuit via the wire.

Abstract: In some implementations, a device may receive, from one or more data sources, information indicating a plurality of data sets, where the plurality of data sets indicate information associated with respective physical locations or online locations. The device may identify a data set, from the plurality of data sets, that indicates information associated with an online location, where the information includes at least one of an entity name, an address, a phone number, a uniform resource locator, an entity identifier, or metadata. The device may parse the data set to identify information for a set of features. The device may analyze the information for the set of features to determine a brand associated with the online location. The device may pair the online location with the brand in the database such that the online location is linked with a first physical location of the brand in the database.

Abstract: Methods and systems disclosed herein describe using machine learning to lock and unlock a device. Machine learning may be trained to recognize one or more features. Once the device has been trained to recognize one or more features, a user may define an unlock condition for the device using the one or more trained features. After defining the unlock condition, the device may be locked by verifying the one or more features that the user defined as the unlock condition using machine learning. When verification is successful, the device may be unlocked and the user allowed to access the device.

Abstract: A device for cooling an electronic component in a cooling fluid immersion environment includes a heatsink, a casing, a micropump, and a first conduit. The heatsink is removably attached to the electronic component and includes baffles that direct the cooling fluid to flow over the electronic component. The cooling fluid absorbs heat from the electronic component. The casing includes an inlet orifice, an outlet orifice, and an internal volume containing the heatsink and the electronic component. The micropump actuates to forcefully direct the cooling fluid from an area surrounding the device, through the inlet orifice, through the series of baffles of the heatsink, and out of the outlet orifice of the casing. Further, the first conduit is connected to the micropump and directs the cooling fluid to the series of baffles of the heatsink within the internal volume of the casing.

Abstract: Systems and methods for automatically detecting, classifying, and processing objects captured in an images or videos are provided. In one embodiment, the system receives an image from an image source and detects one or more objects in the image. The system performs a high-level classification of the one or more objects in the image. The system performs a specific classification of the one or more objects, determines a price of the one or more objects, and generates a pricing report comprising a price of the one or more objects. In another embodiment, the system captures at least one image or video frame and classifies an object present in the image or video frame using a neural network. The system adds the classified object and an assigned object code to an inventory and processes the inventory to assign the classified object a price.

Abstract: The disclosure relates to determining dynamic scores associated with social tokens for senders and receivers in those events and using those scores to assess whether to allow or reject the event. Methods may include receiving data associated with a sender including a social token associated with the sender, receiving account data associated with a plurality of accounts maintained at the entities, generating a health score for the social token associated with the sender, receiving data associated with an event initiated by the sender using the P2P account associated with the sender, generating a health score for the social token associated with the receiver, determining a health score for the event, determining that the event should be approved or declined based on the health score for the event, and if declined locking the application at the network device associated with the sender.

Abstract: Disclosed are systems and methods for effectively managing user access rights, including processing and revalidating user access requests. A computer-implemented method determines an entitlement risk score, a user risk score, and an overall risk score associated with a user access request based on criticality models and metadata. Based on the overall risk score, a criticality level is assigned to the user access request, and the computer-implemented method processes the user access request. The user access request is either automatically or manually processed based on the assigned criticality level. A plurality of user access requests requiring manual processing are processed based on an external input, and the plurality of user access requests are prioritized based on their respective assigned criticality levels and processing histories.

Abstract: Aspects of the subject technology relate to systems and methods of refining manually selected values of physical parameters associated with a manual dip of an azimuthal wellbore image. The method includes perturbating one or more of the manually selected values, selecting a portion of the azimuthal wellbore image based in part on the perturbated values, and performing a Hough transform of the selected portion of the azimuthal wellbore image into a parameter space comprising a plurality of points, thereby determining respective cumulative counts for one or more of the plurality of points in the parameter space. The method also includes selecting a point based in part on its respective cumulative count, calculating refined values of the physical parameters associated with the selected point, and returning the refined values as the physical parameters of a refined dip.

Abstract: The disclosure relates generally to methods and systems for real-time configuration of asset licensing framework for generating non-fungible tokens. Current marketplaces do not safeguard the users with respect to IP rights or copyright infringement since there is no licensing agreement signed at a point of sale of NFTs. The present disclosure provides a standard framework which comprises a configurable set of licensing options that need to be agreed by the transacting parties at the time of NFT sale. These options are embedded into the NFTs in the form of smart contracts, which automatically get executed based on specific events such as secondary sale. The standard license terms are created by administrator that contains both tangible and intangible terms and clauses. The tangible clauses are enforced by smart contracts, whilst intangible clauses can be customized when user is onboarded onto marketplace which can be further extended during each NFT sale.

Abstract: State of art techniques using moderate sized Language Models (LMs) for text classification need fine-tuning or in-context learning. A method and system providing a two-step classification using moderate-sized (#params?2.7B) causal LM (Gen AI) is disclosed. Firstly, for a text instance to be classified, a set of perplexity and log-likelihood based features are obtained from an LM. Further, a light-weight classifier is trained in the second step to predict the final label. The system enables a new way of exploiting the available labelled instances, in addition to the existing ways like fine-tuning LMs or in-context learning. It neither needs any parameter updates in LMs like fine-tuning nor it is restricted by the number of training examples to be provided in the prompt like in-context learning. The key advantages of the disclosed system are explainability through most suitable key phrases and its applicability in resource poor environment.

Abstract: A method of training an artificial neural network to predict gene expression for a patient having a pathological condition is provided. The method comprises obtaining a set of histopathological images from a database of patients having the pathological condition, collecting from the database a set of gene expression profiles corresponding to the histopathological images, identifying in the set of gene expression profiles a set of considered genes, selecting a training subset of the considered genes comprising genes characterized by similar median gene expression values, and training the neural network to predict a gene expression value of an input histopathological image, using gene expression profiles of the genes in the training subset simultaneously.

Abstract: Described herein are systems and techniques for improving an accuracy of determinations made using data sensed in a wellbore or in a laboratory. Nuclear magnetic resonance (NMR) sensing devices may be used to collect data in a wellbore or lab. NMR sensing devices include a magnet (e.g., a permanent magnet or electromagnet) that provides a magnetic field that aligns the spins of protons/nuclei in substances near the NMR sensing device. The magnetic field strength provided by the magnet of the NMR sensing device affects the sensitivity of the NMR sensing device and affects frequencies that the NMR sensing device effectively uses when the NMR sensing device operates. Systems and techniques of the present disclosure may measure concentrations of lithium in brine deposits when identifying particular brine deposits that include sufficient lithium concentrations to justify extracting lithium from those particular brine deposits.

Abstract: A system for improving grid reliability in remote or isolated locations includes a battery tripping module, the battery tripping module determining whether a microgrid zone may be formed upon a fault. The system further includes a microgrid formation module, the microgrid formation module forming a microgrid zone, the microgrid zone connected to a BESS (battery electric storage system), the BESS providing power to the microgrid.

Abstract: Methods and systems for parallel processing of batch requests using a plurality of event-specific processing streams. For example, the system may receive a plurality of requests for batch processing at a preset time. The system may process, with a scheduler service, a first event string of a first request of the plurality of requests. The system may direct the first request to a first processing service for processing within the preset time based on the first event identifier, wherein the first processing service processes the first request with a first processing service configuration file, wherein the first processing service configuration file is specific to the first processing service.

Abstract: A method for multipartite relay may include receiving, from a multipartite relay system, a multipartite relay request. The multipartite relay request may include relay data. The relay data may include a single relay amount and an allocation policy. The method may further include requesting a relay of the single relay amount. The method may further include receiving the relay of the single relay amount. The method may further include, in response to receiving the relay of the single relay amount, splitting the single relay amount into a first apportioned amount and a second apportioned amount according to the allocation policy of the multipartite relay request. The method may further include transmitting the first apportioned amount into a first recipient system of a first recipient and the second apportioned amount into a second recipient system of a second recipient.

Abstract: A system and method perform forensic analysis of electronic data using scoring. The system comprises a processor, a memory, a metric collection module to collect a plurality of metrics of received data, an analysis module to generate a measure of surprise from the metrics and to generate a plurality of scores using the measure of surprise, a detection module to detect problematic data among the received data using the plurality of scores, and a remediation module configured to remediate the problematic data. A display displays the plurality of scores in a column-based visualization sorted using a predetermined visualization selection. The method implements the system.

Abstract: Compounds having useful cannabinoid receptor modulating properties. The compounds may be useful in treating complications arising from metabolic, inflammatory and fibrotic disorders.

Abstract: Aspects of the subject disclosure may include, for example, receiving network information about physical parameters of a fronthaul network of a mobility network, the fronthaul network including multiple network elements including optical communication elements of a radio access network of the mobility network, forming collected network information, receiving from a deep learning (DL) model, operating parameters for a channel in the fronthaul network, the operating parameters determined by the DL model for improved spectral efficiency in the channel between a first element and a second element in the fronthaul network, and launching the channel between the first element and the second element, wherein the launching the channel is according to the operating parameters. Other embodiments are disclosed.