Abstract: Some embodiments relate to methods, systems, and frameworks for data analytics using machine learning, such as methods and systems for preprocessing of biomedical data, using machine learning, for input to a predictive model. The method may include receiving data from a data source, using at least one machine learning (ML) algorithm from a plurality of ML algorithms to obtain at least one combination of preprocessing steps, and computing an accuracy score for each of the at least one combination based on accuracy of prediction of the predictive model. The method may further include using at least one ML algorithm to optimize the feature selection of the predictive model, combining a plurality of datasets into a single dataset, and using a parallel computing network to provide a framework for executing such predictive model.

Abstract: A secure inference over Deep Neural Networks (DNNs) using secure two-party computation to perform privacy-preserving machine learning. The secure inference uses a particular type of comparison that can be used as a building block for various layers in the DNN including, for example, ReLU activations and divisions. The comparison securely computes a Boolean share of a bit representing whether input value x is less than input value y, where x is held by a user of the DNN, and where y is held by a provider of the DNN. Each party computing system parses their input into leaf strings of multiple bits. This is much more efficient than if the leaf strings were individual bits. Accordingly, the secure inference described herein is more readily adapted for using in complex DNNs.

Abstract: The present invention introduces an innovative approach to cloud-based image storage that leverages grid-based segmentation and selective hashing to efficiently reduce storage space. By identifying and interning common image segments—while preserving unique elements such as faces—the system achieves a balance between storage efficiency and computational overhead. This method not only minimizes redundancy but also optimizes resource utilization, making it a compelling solution for large-scale cloud storage providers. The invention represents a significant advancement over traditional deduplication techniques, offering improved storage management in environments where repeated image content is prevalent.

Abstract: The present invention relates to event-driven ERP integration framework for real-time data synchronization. The Event-Driven ERP Integration Framework revolutionizes enterprise system integrations by shifting from batch-based processing to a real-time event-driven model. This approach ensures instant data synchronization, enhanced scalability, and improved fault tolerance, significantly reducing operational overhead and inefficiencies. By leveraging message-driven architectures and AI-enhanced anomaly detection, the framework empowers enterprises with agile, responsive, and intelligent ERP integrations, positioning them for sustained operational excellence in dynamic business environments.

Abstract: An example system, e.g., a system on a chip (SoC), includes first and second domains having first and second processors, respectively. The second processor is part of a processing subsystem in the second domain. The first processor provides an instruction to the second processor, which executes the instruction to configure the processing subsystem to operate in a mode specified by the instruction. In response to the processing subsystem being configured to operate in the specified mode, isolation circuitry of the system is configured to provide a level of isolation between the first domain and the second domain based on the specified mode.

Abstract: A verifiable identity map maintains identities and public keys associated with the identities. The map is maintained by a ledger database that provides tamper-resistant/evident capabilities for tables (comprising the map) thereof. For instance, when a materialized view of the database is generated, the database provides a digest representative of a state thereof to computing devices that access the map for the keys. When the database receives a request from a device to access the map, the digest is received along therewith. The database is validated based on the digest to determine whether the database has been tampered with since the provision of the digest. Responsive to a successful validation, the database provides access in accordance with the request. When a key in the map is updated, the database subsequently generates a new digest, which is provided to the computing device.

Abstract: The present invention relates to a Digital Green Score (DGS) system and a framework that evaluates the environmental impact of digital content usage, infrastructure management, and resource consumption. The DGS is generated at the individual level, aggregating scores across all applications and systems the person is responsible for. The organization-wide DGS is an aggregation of all individual scores, ensuring comprehensive sustainability management. The present invention aims to establish a global standard for measuring digital sustainability and ensuring a responsible, eco-friendly future for digital ecosystems.

Abstract: Embodiments are directed to a method for determining relationships between genes, phenotypes, and diseases that includes extracting a first set of biological data containing gene data, phenotype data, and disease data from a database, extracting a second set of biological data from a database, creating nodes on a network pertaining to each individual gene, phenotype, and disease data extracted from one of a database and documents, training multiple machine learning (ML) algorithms on a set of extracted data with matching empirical results, using at least one of the ML algorithms and the second set of biological data for determining relationships between the nodes, creating at least one of a gene-disease association score, gene-phenotype association score, and disease-phenotype association score for the relationships based on the relative association of the nodes, and displaying at least one of the association scores to a user.

Abstract: An apparatus includes a first input power bus having a first input terminal coupled to a first input power source, a first output terminal, and a first group of line diodes between the first input terminal and the first output terminal. The apparatus also includes a second input power bus having a second input terminal coupled to a second input power source, a second output terminal coupled to the first output terminal of the first input power bus, and a second group of line diodes between the second input terminal and the second output terminal. The first output terminal and the second output terminal are coupled to provide power to a load.

Abstract: A deployment orchestrator is provided that manages package deployments at different hierarchical levels. Each hierarchical level is associated with a particular type of resource object. The deployment orchestrator creates different of resource objects, each associated with a different hierarchical level and updates instances of the different resource objects based on information related to a package that is to be deployed. The deployment orchestrator performs processing associated with deploying the package at the hierarchical level based on information stored in the instances of the resource objects associated with the hierarchical level e.g., information related to a package that is to be deployed.

Abstract: A system, e.g., a system on a chip (SoC) includes a first domain including a first processor configured to boot the system; a second domain including a processing subsystem having a second processor; and isolation circuitry between the first domain and the second domain During boot-up of the system, the first processor provides code to the second domain. When the code is executed by the second processor, it configures the processing subsystem as either a safety domain or as a general-purpose processing domain. The safety domain may an external safety domain or an internal safety domain.

Abstract: Techniques are disclosed relating to the execution of machine learning models on client devices, particularly in the context of transaction risk evaluation. This reduces computational burden on server systems. In various embodiments, a server system may receive, from a client device, a request to perform a first operation and select a first machine learning model, from a set of machine learning models, to send to the client device. In some embodiments the first machine learning model is executable, by the client device, to generate model output data for the first operation based on one or more encrypted input data values that are encrypted with a cryptographic key inaccessible to the client device. The server system may send the first machine learning model to the client device and then receive, from the client device, a response message that indicates whether the first operation is authorized based on the model output data.

Abstract: The invention relates to a method for training machine learning models, having the steps of: detecting data in the form of time series data using one or more computers, said data being obtained by means of one or more measuring devices (60-62), in each case in the form of a sensor for measuring a physical variable; receiving multiple classification data units relating to the data using the one or more computers; receiving a selected part of the data using the one or more computers for each of the classification data units; and training multiple machine learning models using the one or more computers, in each case on the basis of at least one of the classification data units and the at least one corresponding selected part of the data, wherein the multiple machine learning models represent multiple instances of the same machine learning model.

Abstract: A method of operating a HVAC system is provided. The method includes receiving a first outdoor airflow and a return airflow in a first heat exchange space of a housing, where the housing comprises an interior wall that divides the housing into the first heat exchange space and a second heat exchange space. The method includes transferring the first outdoor airflow and a first portion of the return airflow across a first heat exchanger in the first heat exchange space to produce a first conditioned airflow that discharges to a target conditioned space. The method includes discharging a second portion of the return airflow from the first heat exchange space as exhaust airflow. The method includes transferring a second outdoor airflow and the portion of the exhaust air across a second heat exchanger in the second heat exchange space.

Abstract: The present disclosure relates to methods and systems that preserve privacy in a secure multi-party computation (MPC) framework in multi-agent reinforcement learning (MARL). The methods and systems use a secure MPC framework that allows for direct computation on encrypted data and enables parties to learn from others while keeping their own information private. The methods and systems provide a learning mechanism that carries out floating point operations in a privacy-preserving manner.

Abstract: In described examples, a method of operating a system on chip (SoC) includes the following steps. Receive a component description that describes connections among circuit elements of the SoC. Receive a list of specified peripherals to activate while the SoC boots. Receive a hint data that describes a bypass mode clock frequency of a phase locked loop (PLL) of the SoC. Determine an operational data in response to the component description, the hint data, and the specified peripherals. The operational data specifies a subset of the circuit elements required to make the specified peripherals operational. Initialize, and limit the initializing, to the subset of the circuit elements.

Abstract: A system and method for enhancing power flow analysis convergence are provided. The method includes receiving a first power system artifact from one or more systems associated with a power system. The first power system artifact includes a plurality of data elements associated with the power system at a given time instant. The method further includes extracting a dataset from an archival system. The dataset is extracted based on one or more appropriate data elements selected from the plurality of data elements. The method further includes determining from the extracted dataset a prediction of initial conditions for the power flow analysis using a machine learning algorithm and assigning target variables of the predicted initial conditions to the first power system artifact. Thereafter, the method includes performing the power flow analysis on the first power system artifact using the assigned target variables.

Abstract: Data diode systems and methods are disclosed herein for enhancing data security. Encrypted data transmitted from a first node (e.g., an entity coupled to a network) is received. The data transmitted is encrypted with a public key associated with a second node (e.g., the node to which to which the encrypted data is transmitted). The encrypted data is decrypted with a private key associated with the second node to generate decrypted data. A determination is made whether a digital signature in the decrypted data corresponds to a ledger entry mapped to the first node in a first set of ledger entries. The first node is verified to be a trusted entity based on the digital signature having been determined to correspond to the ledger entry. Based on the verification, the transmission of the encrypted data from the first node is determined to be a permissible data transmission.

Abstract: An electric-to-hydraulic conversion machine includes an axial flux electric motor and a hydraulic pump. The motor includes a spindle, at least one rotor configured to rotate about the spindle, and at least one stator configured to drive rotation of the at least one rotor about the spindle. The hydraulic pump includes a piston block having a plurality of cylinders, a plurality of pistons each supported in one of the cylinders of the piston block, and a cam ring. The piston block is attached to the at least one rotor and is configured to rotate about the spindle with rotation of the at least one rotor. The cam ring is configured to radially drive the pistons during rotation of the piston block about the spindle, which drives a fluid flow.

Abstract: An entity is enabled to access encrypted resources in response to verifying access criteria of a region-based security policy is met. For example, a resource request to access an encrypted resource is received from an entity. A determination that the encrypted resource is assigned to a first region and is protected by a region-based security policy is made. A proof of a region attribute indicating that the entity possesses the region attribute is received from the entity, the region attribute indicates the entity is associated with the first region. An encrypted version of the region attribute is obtained from a ledger database. The resource request is validated based at least on the encrypted attribute and the proof of the region attribute. A verification is made that an access criteria of the region-based security policy is met. The entity is provided access to the encrypted resource.