Abstract: This disclosure is directed to a generalizable machine learning model production environment and system with a defense mechanism that facilitates safe execution of machine learning models in production by effectively detecting potential known and new adversarial attacks. The disclosed exemplary systems and architectures gather data from the online execution of the machine learning models and communicate with an on-demand pipelines for further inspection and/or correction of vulnerabilities in the production machine learning model to the detected attacks. These systems and architectures provide an automatable process for continuous monitoring of model performance and correction of the production machine learning model to guard against current and future adversarial attacks.

Abstract: A method and system for evaluating and enhancing a prompt for use by a generative artificial intelligence processing model are disclosed. The method may include obtaining a prompt representing a natural language text for use by a generative artificial intelligence processing model, obtaining a prompt classifier trained to evaluate a classification of the prompt, and inputting the prompt to the prompt classifier to generate a classification of the prompt. The method may further include identifying an intent underlying the prompt and detecting an implicit constraint for the prompt based on the intent. The method may further include transforming the intent in the prompt into a constraint-enhanced intent based on the implicit constraint, generating an enhanced prompt based on the constraint-enhanced intent, and outputting the enhanced prompt for the generative artificial intelligence processing model.

Abstract: A system for governing user privacy data may include detecting a dark data source in a data ecosystem including data sources storing user privacy data. The system may further include identifying user information from the data ecosystem according to a lineage of datasets and scanning the dark data source by executing a natural language processing engine to identify existence of and content of user privacy data in the dark data source based on the user information. The system may further include selecting a target user privacy data compliance policy from user privacy data compliance policies based on a geographical region associated with the dark data source and detecting non-compliance in protecting the user privacy data in the dark data source based on the target user privacy data compliance policy. The system may further include, in response to the non-compliance, processing the user privacy data to eliminate the non-compliance.

Abstract: In some implementations, an application programming interfaces (API) manager may receive, at a set of artificial intelligence (AI) APIs, a set of inputs from a set of on-site devices. Accordingly, the API manager may route the set of inputs to a corresponding set of remote servers and may receive, from at least one server of the corresponding set of remote servers, at least one response based on at least one input, from the set of inputs, routed to the at least one server. The API manager may transmit the at least one response to a corresponding device from the set of on-site devices. Further, the API manager may modify at least one API, of the set of AI APIs, based on a traffic pattern associated with the set of inputs and the at least one response.

Abstract: The present disclosure describes a method of controlling a manufacturing system using multivariate time series and includes storing recording data as a plurality of time series, each time series having a first recorded value and a final recorded value, interpolating, within a first time window, missing values in the plurality of time series using a Bayesian model, the missing values falling between a first and an end time of the respective time series, storing the interpolated values as prediction data, each interpolated value including an uncertainty, loading recorded data of a second time window, loading prediction data of the second time window, predicting, using the Bayesian model, values for each time series that is absent recorded data and prediction data, storing the predicted values, each prediction value including an uncertainty, and adjusting a device that generates the recorded data based on the prediction values within the second time window.

Abstract: A system may provide blockchain-based storage node control for unified proof-of-storage interaction for off-chain data chunk storage. The system may include verification circuitry that may provide a challenge to a storage node to audit the storage status of a data chunk. The verification circuitry may obtain identities for the storage node and data chunk based on blockchain records. The verification circuitry may obtain a challenge answer from the storage node in response. The verification circuitry may analyze the challenge answer in view of a challenge key to confirm whether the storage node has possession of the data chunk. When the storage node has possession of the data chunk chain circuitry may add a verification record to the blockchain.

Abstract: In some implementations, a system may receive, from an analysis system, a data stream associated with monitoring an environment. The system may process, using a first model, the data stream to detect an object associated with a service provided by an application. The system may analyze a monitoring configuration of the analysis system. The system may determine, based on processing the data stream, service data associated with providing the service in association with the object. The system may configure, based on the monitoring configuration, an edge device to access the data stream and host a second model to process the service data according to the data stream. The system may receive, from the edge device, a service output associated with providing the service in association with the object. The system may perform an action associated with the service output and the object.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support automated optimization of product management systems. In embodiments, automated optimization of component tools of the product management system is provided by automated evaluation and optimization of associated rules and metadata. In embodiments, metadata and rules may be associated to each other by an assessment engine. A recommendation engine may then identify non-compliant metadata, may determine a condition of the rule, and/or may generate recommendations for the rules based on the non-compliant metadata. Automated optimization of the product management system may include automated creation and mapping of decomposition relationships between commercial products and technical products. In embodiments, input data may be parsed into a set of object with unique attribute fields, which may then be validated.

Abstract: A method and system for training a federated learning model are disclosed. The method may include receiving the global federated learning model from the server via the client. The client may control remote computing resources. The method may further include identifying a spare computing instance from the remote computing resources and determine a threshold training load for training the global federated learning model based on a training load assigned to the client. The method may further include, in response to a processing capacity of the spare computing instance being sufficient to process the threshold training load, offloading the threshold training load to the spare computing instance and training the global federated learning model on the spare computing instance with the training dataset stored in a data source maintained by the client.

Abstract: A device may configure a plurality of time dimensions of a forecast model to forecast transactions. The device may identify, using the forecast model and transaction information associated with forecasting a transaction, historical transaction data that includes timing information associated with executions of a set of previous transactions that are associated with the transaction. The device may generate, using the forecast model and the plurality of time dimensions, a plurality of clusters associated with the set of previous transactions. The device may determine, using the forecast model, similarity scores for the plurality of clusters that correspond to individual probabilities that the set of previous transactions occurred according to the plurality of time dimensions. The device may select, based on the similarity scores, a time dimension for a forecast of the transaction. The device may generate the forecast of the transaction based on the time dimension.

Abstract: Methods, systems, and computer-readable storage media for receiving data representative of two or more AAGs, providing an identifier for each element of each of the two or more AAGs, each identifier being unique within a respective AAG, at least one identifier being non-unique between the two or more AAGs, determining an attribute value for each element of each of the two or more AAGs, storing attribute value to element mappings in an attribute dictionary, providing a differenced AAG based on the attribute value to element mappings in the attribute dictionary, determining a set of remedial actions at least partially based on the differenced AAG, and executing one or more remedial actions in the set of remedial actions to reduce a cyber security risk to the enterprise network.

Abstract: A new product development system may provide, to one or more client devices and server devices, one or more tools to enable development of a new product. The new product development system may receive current project data identifying a current project for developing the new product, historical project data identifying historical projects for developing historical products, and industry data associated with the new product. The new product development system may process the current project data, the historical project data, and the industry data, with a machine learning model, to identify one or more variables for a predictive model and may process the one or more variables, with the predictive model, to predict a predicted success rate of the current project. The new product development system may perform one or more actions based on the predicted success rate of the current project.

Abstract: Embodiments of the present disclosure leverage near field communication (NFC) technology to provide dynamic and interactive monitoring of an environment. NFC devices may be used to check items in and out of a storage facility, obtain readings from instruments or other machinery present in the environment (e.g., to perform tests on the items, etc.), track movement of users and items within the environment, and to prompt users with information about the environment, such as information about the instruments or machinery currency being used to perform operations with respect to one or more items checked out to the user. Additionally, the NFC device may be used to configure the instruments with appropriate settings for the particular item(s) for which the instrument is currently being used or for other purposes.

Abstract: A method and system for generating structured semantic annotations from unstructured documents are disclosed. The method may include obtaining a natural language processing model trained to extract unstructured data from an unstructured document and output structured data based on a generic document ontology. The method may further include inputting a plurality of unstructured documents into the natural language processing model to obtain generic structured annotations of the unstructured documents. The method may further include aligning the generic structured annotations to a target ontology specifying a use case to generate target structured annotations for the use case. The target structured annotations may include entities and relations between the entities. The method may further include generating a knowledge representation for the use case based on the target structured annotations and generic structured annotations.

Abstract: Methods and systems for dynamic dashboard administration are presented. Embodiments include displaying a dashboard through a graphical user interface (‘GUI’); receiving user instructions to manipulate one or more elements of the dynamic dashboard; capturing the manipulation of the one or more GUI elements; parsing, according to a gesture taxonomy and gesture ontology, the captured manipulation into one or more gesture triples; storing the gesture triples in an enterprise knowledge graph of a semantic graph database; identifying, in dependence upon the gesture triples, dashboard insights; updating the dashboard in accordance with the dashboard insights.

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: In some implementations, a device may receive project management data including information related to project elements associated with one or more projects and associations among the project elements associated with the one or more projects. The device may generate one or more data structures representing the project elements associated with the one or more projects and the associations among the project elements associated with the one or more projects. The device may receive a request to visualize project data, wherein the request may indicate a view to use in order to visualize the project data and/or specify criteria related to the project data to be visualized in the view. The device may search the one or more data structures to obtain the project data to be visualized based on the criteria specified in the request. The device may render the project data using the view indicated in the request.

Abstract: A device may receive demand data associated with a product or a service, multiple forecasting models, and multiple cost functions, and may identify primary parameters for the multiple models based on the demand data. The device may utilize a model to rank the multiple forecasting models and the multiple cost functions based on the primary parameters, and may select optimum primary parameters based on ranking the multiple forecasting models and the multiple cost functions. The device may identify secondary parameters for the multiple forecasting models based on the demand data and the optimum primary parameters. The device may select optimum secondary parameters based on ranking the multiple forecasting models and the multiple cost functions, and may select a forecasting model, from the multiple forecasting models, based on the optimum primary parameters and the optimum secondary parameters. The device may perform one or more actions based on selecting the forecasting model.

Abstract: This disclosure describes example implementations for generating context-dependent embedding vectors of words in the multi-dimensional embedding space based on a generic and non-domain-specific pretrained word embeddings. Such an implementation requires no domain specific training corpus and is capable of generating context-dependent embedding vectors of multi-semantic words using a few contextual texts. Such an implementation thus provides an efficient way to generate a library of multiple domain specific embedding vectors for multi-semantic words without any domain-specific training process. Other example embodiments further apply the principles of the context-dependent word embedding generation to a text-to-image application.

Abstract: A device may receive work item data identifying work items associated with requirements from different tools of a project and may perform data cleansing to remove and/or modify particular words from the work item data and to generate cleansed work item data. The device may perform natural language processing on the cleansed work item data to identify synonyms for words in the cleansed work item data and may replace abbreviations in the cleansed work item data with full form text to generate final work item data. The device may identify keywords in the final work item data and may process the final work item data, the synonyms, and the keywords, with a machine learning model, to identify mappings between work items of the final work item data and to determine a confidence score for the mappings. The device may perform actions based on the mappings and the confidence score.