Systems, methods, and apparatuses may implement a degradation-based evaluator to verify performance of evaluation metrics. A degradation-based evaluator may iteratively degrade a candidate text to generate a series of candidate texts, then submit the series of candidate text to a target evaluator. The target evaluator may evaluate the series of candidate texts with respect to a reference text to generate a series of performance metrics, then degradation-based evaluator may evaluate the series of performance metrics to verify the performance of the target evaluator.
Mechanisms are provided for implementing a shared memory segment-based (SMS) cache that improves performance of isolated application instances with regard to startup times and horizontal scaling. A plurality of application instances are executed in a cloud computing environment runtime, where each instance executes in an isolated manner from each other instance. A shared memory segment (SMS) for caching data used by the plurality of application instances is allocated, and each application instance is attached to the SMS. A first application instance, via a first custom call to the cloud computing system runtime, stores data to a cache implemented in the shared memory segment. A second application instance retrieves, via a second custom call to the cloud computing system runtime, cached data from the cache without accessing an original data source of the cached data.
A system provides a delegate credential to access a target resource based on one or more access policies associated with a delegate principal. The system receives a credential request for the delegate credential from a computing entity associated with a recipient principal. The system transmits an approval request to an approval service associated with the target resource for the approval service to approve issuance of the delegate credential to the recipient principal. The system receives an approval confirmation from the approval service and generates the delegate credential responsive to receiving the approval confirmation. The system transmits the delegate credential to a computing entity associated with the recipient principal. The computing entity accesses the target resource by presenting the delegate credential to a resource service associated with the target resource.
Operations of a certificate bundle validation service may include receiving a first certificate bundle that includes a first set of one or more digital certificates, and a digital signature, associated with the first certificate bundle; determining, using a public key of an asymmetric key pair associated with a second set of one or more digital certificates, that the digital signature is generated using a private key of the asymmetric key pair; and responsive to determining that the digital signature is generated using the private key, storing the first certificate bundle in a certificate repository as a trusted certificate bundle.
H04L 9/32 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité comprenant des moyens pour vérifier l'identité ou l'autorisation d'un utilisateur du système
Systems, methods, and other embodiments associated with design of infrastructures and other solutions by LLM using a variable syntax for graphs are described. In one embodiment, a method includes accessing requirements for a solution that are in human language. The method includes generating a prompt based at least in part on the requirements. The prompt is configured to cause a large language model to generate a solution topology in a graph representation language. The graph representation language may vary in syntax in response to a change to the requirements. And, the method includes, in response to the prompt, generating the graph representation language, and the solution topology in the graph representation language with the large language model.
Training large language models using masked language modeling is computationally expensive and data inefficient, as it requires multiple forward passes to learn from individual masked tokens in a sequence. A method for training language models is introduced using unmasked positive data and permuted negative data to address these limitations. Positive data includes grammatical, well-formed sequences, while negative data includes randomly or validated permuted versions of the positive data, designed to disrupt linguistic structure. The method trains the model to assign high probabilities to positive data and low probabilities to negative data, using anti-modeling techniques to prevent the model from adopting undesirable shortcuts, such as copying input tokens. This approach enables the language model to learn meaningful linguistic representations in fewer forward passes, significantly improving data and computational efficiency while maintaining robust generalization capabilities.
Described is an improved approach to address software incompatibility issues arising from software upgrades. One way to address this is to leverage a Highly Available IP (HAIP) failover mechanisms, with the automated migration of IP addresses between broadcast domains to ensure all nodes operate within a single, compatible environment during software upgrades or migrations, followed by a seamless failback to the original configuration upon successful completion.
H04L 41/0663 - Gestion des fautes, des événements, des alarmes ou des notifications en utilisant la reprise sur incident de réseau en réalisant des actions prédéfinies par la planification du basculement, p. ex. en passant à des éléments de réseau de secours
H04L 41/082 - Réglages de configuration caractérisés par les conditions déclenchant un changement de paramètres la condition étant des mises à jour ou des mises à niveau des fonctionnalités réseau
8.
SCALING LOG DATA PROCESSING BEYOND A LOG INDEX WITH DYNAMIC HELPERS SUPPORTING THE SAME LOG INDEX
Techniques for processing log data items for a log index using multiple queues for processing and storing information from the log data items to multiple different storage systems are provided. A log processing system accesses a queue of log data items mapped to a log index, and detects that log data items are being added to the queue faster than processing of log data items from the queue. Based at least in part on the detecting, the log processing system activates a separate queue for the same log index. The separate queue is supported by other log processing instance(s) that are configured to process the separate queue's log data items and store sets of resulting information to a second storage system. The log processing system also stores an indication that data having characteristic(s) is stored across both the first storage system and the second storage system.
A system and method for use of spreadsheets integrated with web services, including support for creation of child rows in child spreadsheets. A client device can interrogate one or more web services to obtain web service metadata that determines one or more capabilities of those web services. This metadata can be used to automatically configure an editable table component in a spreadsheet. The table component can be used to retrieve data, edit that data in a spreadsheet, and upload changes back to a web service, or provide validations, recommendations, or other functionalities. A user interface allows a user to create child layouts, including inputting a number of child rows to be created for selected parent layout or rows. The system then activates a child layout with appended child rows, including automatically selecting the first editable cell in the first child row and populating ancestor values indicative of the hierarchical relationship.
G06F 40/18 - Édition, p. ex. insertion ou suppression de tableauxÉdition, p. ex. insertion ou suppression utilisant des lignes réglées de tableurs
G06F 3/0482 - Interaction avec des listes d’éléments sélectionnables, p. ex. des menus
G06F 3/04845 - Techniques d’interaction fondées sur les interfaces utilisateur graphiques [GUI] pour la commande de fonctions ou d’opérations spécifiques, p. ex. sélection ou transformation d’un objet, d’une image ou d’un élément de texte affiché, détermination d’une valeur de paramètre ou sélection d’une plage de valeurs pour la transformation d’images, p. ex. glissement, rotation, agrandissement ou changement de couleur
A method implements formal specification assisted program repair. The method involves processing buggy code to generate a specification from the buggy code using a specification generation prompt. The method further involves verifying the specification to generate a verified specification. The method further involves injecting the verified specification into a repair prompt. The method further involves generating a candidate patch for a buggy method in the buggy code using the repair prompt.
A deployment system is disclosed that reduces or eliminates redundant deployments. For a particular artifact image to be deployed to a computing environment, the deployment system initiates the deployment only if (a) there has been no previous deployment of the particular artifact image to the computing environment, or (b) the version of the particular artifact image to be deployed is different from the previous version of the particular artifact image that was deployed to the computing environment and/or the computing environment has changed since that previous deployment. Different techniques are used to determine if the version of the particular artifact image to be deployed is same as or different from the previously deployed version of the particular artifact image, and also to determine if the computing environment has changed since that previous deployment.
Techniques for distributed edge node task offloading management are disclosed. An edge server implementing an artificial intelligence (AI) model receives requests from devices to process tasks using the AI model. The edge server analyzes local edge server attributes, remote node attributes, and network attributes to determine whether to offload the task to another edge node. The edge node determines if it has sufficient processing capacity to perform the task in a specified timeframe. The edge node further determines if another edge node has sufficient processing capacity to perform the task. The edge node further determines whether to offload the task based on pending tasks on the other edge node, based on latency values associated with offloading the task, and based on energy consumption associated with offloading the task. If the edge node determines the offloading criteria are met, the edge node offloads the task to another edge node.
H04L 67/1001 - Protocoles dans lesquels une application est distribuée parmi les nœuds du réseau pour accéder à un serveur parmi une pluralité de serveurs répliqués
Techniques for detecting and tracking disposable email addresses are detected. A first email associated with a first domain and a second email associated with a second domain are received. The first domain and the second domain may be different. An analysis is performed on a plurality of DNS (domain name system) records comprising at least a first DNS record of a first type associated with the first email and a second DNS record of a second type associated with the second email. The first type and the second type are different. Whether a correlation of information amongst the plurality of DNS records satisfies a security criterion is determined. In an example, responsive to (1) determining that the first domain has been flagged as malicious and (2) determining that the correlation of information amongst the plurality of DNS records satisfies the security criterion, the second domain is flagged as malicious.
Techniques for generating and resolving compact enum references are disclosed. A virtual machine (VM) detects an instruction to store a reference to an enum value of an enum type. Responsive to encountering the instruction to store the reference to the enum value, the VM determines an ordinal position of the enum value in an enum lookup array that includes a set of possible values of the enum type. The VM stores the ordinal position of the enum value as the reference to the enum value of the enum type.
Techniques are disclosed herein for an AI-powered coding assistant with generative model for code development. In one aspect, a method is disclosed that includes: accessing customization information comprising code examples and code instructions associated with one or more code tasks, generating, by a generative model, a set of training examples based on the customization information and the one or more code tasks. The set of training examples is generated in an iterative process that includes: (i) generating, based on the customization information, a training prompt, (ii) generating, based on the training prompt, a training completion, and (iii) associating the training prompt with the training completion to generate a training example of the set of training examples. The set of training examples are then used to fine-tune a pre-trained machine learning model to generate a fine-tuned machine learning model.
Techniques for distributed edge node task offloading management are disclosed. An edge server implementing an artificial intelligence (AI) model receives requests from devices to process tasks using the AI model. The edge server analyzes local edge server attributes, remote node attributes, and network attributes to determine whether to offload the task to another edge node. The edge node determines if it has sufficient processing capacity to perform the task in a specified timeframe. The edge node further determines if another edge node has sufficient processing capacity to perform the task. The edge node further determines whether to offload the task based on pending tasks on the other edge node, based on latency values associated with offloading the task, and based on energy consumption associated with offloading the task. If the edge node determines the offloading criteria are met, the edge node offloads the task to another edge node.
A system identifies a secure entry point for accessing a target resource of a virtual cloud network. The system determines a mapping that maps the secure entry point to a particular addressable network entity of a plurality of addressable network entities. Based on the mapping, the system selects the particular addressable network entity as a destination for requests associated with the target resource. The system transmits a network address corresponding to the particular addressable network entity to a computing entity as a destination address for the requests associated with the target resource. The computing entity transmits the requests associated with the target resource to the network address corresponding to the particular addressable network entity, and the addressable network entity forwards the requests to the secure entry point.
H04L 9/36 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité avec des moyens pour détecter des caractères non destinés à la transmission
H04L 9/38 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité le chiffrement étant effectué par un appareil mécanique, p. ex. cames rotatives, interrupteurs, perforatrices de bande à clavier de clé
Systems, methods, and other embodiments associated with detecting security vulnerabilities in heap memory usage are described. In one embodiment, a runtime data model is based on a petri net configured to track heap memory usage in a target system. The runtime data model tracks heap memory allocation and deallocation function calls. The runtime data model is simulated in the petri net with runtime data the target system, wherein the runtime data model associates the heap memory allocation functions with corresponding heap memory deallocation functions in a visual graphical format. Security vulnerabilities are detected by the model, for example, where a specific heap memory allocation function does not have a corresponding heap memory deallocation function. An alert may be generated that identifies the security vulnerabilities including the associated specific heap memory allocation function that does not have the corresponding heap memory deallocation function.
G06F 21/57 - Certification ou préservation de plates-formes informatiques fiables, p. ex. démarrages ou arrêts sécurisés, suivis de version, contrôles de logiciel système, mises à jour sécurisées ou évaluation de vulnérabilité
19.
AI-POWERED CODING ASSISTANT WITH GENERATIVE MODEL FOR CODE DEVELOPMENT
Techniques are disclosed herein for an AI-powered coding assistant with generative model for code development. In one aspect, a method is disclosed that includes: accessing customization information comprising code examples and code instructions associated with one or more code tasks, generating, by a generative model, a set of training examples based on the customization information and the one or more code tasks. The set of training examples is generated in an iterative process that includes: (i) generating, based on the customization information, a training prompt, (ii) generating, based on the training prompt, a training completion, and (iii) associating the training prompt with the training completion to generate a training example of the set of training examples. The set of training examples are then used to fine-tune a pre-trained machine learning model to generate a fine-tuned machine learning model.
In accordance with various embodiments, described herein is a system and method for use of spreadsheets integrated with web services including support for action item management. A client device can interrogate one or more web services to obtain web service metadata that determines one or more capabilities of those web services. This metadata can be used to automatically configure an editable table component in a spreadsheet. The table component can then be used to retrieve data, edit that data in a spreadsheet, and upload changes back to a web service, or provide validations, recommendations, or other functionalities. An action items management component assists the workflow of navigation, inspection, correction, and task-tracking of action items associated with spreadsheets having large amounts of cells, to enable focus on those cells necessitating user review or action, for example to address upload results, data validations, or other issues associated with those cells.
Embodiments provide cloud based health insurance enrollment for a user. Embodiments determine that open health insurance enrollment is available for the user and, using a pre-trained large language learning machine learning (“ML”) model, automatically presents to the user a plurality of available health insurance options based on a current health insurance or qualifications for insurance. Embodiments, using the ML model, automatically analyze the available health insurance options and, using a generative artificial intelligence (“AI”) model, automatically answer questions from the user regarding the available health insurance options.
A clustering model is used to partition a set of historical opportunity profiles representing historical opportunities into a plurality of opportunity profile clusters. A target opportunity profile that includes a plurality of target-opportunity parameters is obtained, and an opportunity profile cluster is identified that meets a matching criterion with the target opportunity profile. The plurality of historical opportunity profiles of the identified opportunity profile cluster are partitioned based on an outcome indicator associated respectively with each of the plurality of historical opportunity profiles, providing a first subset of historical opportunity profiles and a second subset of historical opportunity profiles. An insight generation model is applied to the target opportunity profile along with the first subset and/or the second subset to generate one or more insights for the target opportunity profile.
A file system stores data elements at a corresponding plurality of offsets. The file system removes a first subset of the data elements while retaining a second subset of the data elements. The file system receives, from an application, a request for a sequentially-first offset that corresponds to a sequentially-first valid data element that is still retained by the file system. The file system determines the sequentially-first offset and transmits the sequentially-first offset to the application. Subsequently, the file system receives, from the application, a request for one or more retained data elements commencing from the sequentially-first valid data element. The request identifies the sequentially-first offset. In response to the request for the one or more retained data elements, the file system transmits to the application at least a portion of the second subset of the data elements commencing from the sequentially-first valid data element.
A control plane of a source cloud environment receives first information included in a first metadata instance and second information included in a second metadata instance. The first information represents a logical organization of resources that is to be created in the source cloud environment, and the second information indicates one or more networking resources that are to be created in the source cloud environment. The control plane of the source cloud environment creates the logical organization of resources associated with the first metadata instance and the one or more networking resources associated with the second metadata instance in the source cloud environment. One or more resources associated with a service provided by the source cloud environment are provisioned via a console associated with the source cloud environment based on the first metadata instance and the second metadata instance.
Techniques for increasing the precision of machine learning models that are trained to detect anomalous data points in a time series. The techniques including methods and systems for training machine learning models offline, using the trained machine learning models to predict anomalies in an online runtime environment, and updating anomaly detection models in the runtime and offline environments. The machine learning models may include a multitask model for predicting one or more anomalous events present in input time series data, and for each identified anomalous event type, predicting a type of machine learning model that is best suited for predicting that anomalous event type. The models may further include model instances selected using the predicted one or more anomalous events present in the input time series data and the predicted anomalous event type and used to predict an anomaly event in the input time series data.
A method can include identifying handwritten elements from a document image. The method can include generating embeddings for the handwritten elements, and the embeddings may represent features extracted from the handwritten elements. The method can include clustering the embeddings into clusters. The method can include processing the document image to identify areas within the document image designated for placement of a signature. The method can include, based upon locations of the areas within the document image, tagging each cluster as a signature cluster or as a non-signature cluster. The method can include outputting indications that a handwritten element is a signature element or is a non-signature element.
G06V 40/30 - Reconnaissance d’auteurLecture et vérification des signatures
G06T 7/70 - Détermination de la position ou de l'orientation des objets ou des caméras
G06V 10/40 - Extraction de caractéristiques d’images ou de vidéos
G06V 30/412 - Analyse de mise en page de documents structurés avec des lignes imprimées ou des zones de saisie, p. ex. de formulaires ou de tableaux d’entreprise
27.
Topic-Based Synchronization Of A Chat Interface With An Information Interface
Techniques for synchronizing the topic of messages in a chat interface with content in an information pane include concurrently displaying a chat interface and an information pane in a GUI and monitoring the content currently displayed within the information pane to determine a set of one or more topics. The system identifies and displays a subset of stored chat messages associated with the set of topics concurrently with the content. When the system detects that the content in the information pane is changed to different content, the system determines a second set of one or more topics corresponding to the changed content and identifies a second subset of the stored chat messages based on the second set of topics. The second subset is displayed in the chat interface concurrently with the second set of content.
H04L 51/04 - Messagerie en temps réel ou quasi en temps réel, p. ex. messagerie instantanée [IM]
H04L 51/02 - Messagerie d'utilisateur à utilisateur dans des réseaux à commutation de paquets, transmise selon des protocoles de stockage et de retransmission ou en temps réel, p. ex. courriel en utilisant des réactions automatiques ou la délégation par l’utilisateur, p. ex. des réponses automatiques ou des messages générés par un agent conversationnel
H04L 51/216 - Gestion de l'historique des conversations, p. ex. regroupement de messages dans des sessions ou des fils de conversation
28.
Managing Resource Usage Monitoring Using Machine Learning and Rules-Based Models
Techniques for improving the performance of a resource usage monitoring system by selectively verifying anomaly predictions are disclosed. A system inputs resource usage data to a machine learning model. The system determines remediation actions to perform based on whether or not the system verifies machine-learning-generated anomaly predictions. The system verifies the anomaly predictions by applying measured resource usage data to a rules-based verification model. If the system verifies a predicted anomaly, the system remediates faults in resource metering devices, such as sensors, memory chips, microcontrollers, and data transmission devices. If the system determines a predicted anomaly is not verified, the system remediates a fault in the machine learning model by generating a training record based on the measured and estimated resource usage values to retrain the machine learning model.
Techniques for validating a credential holder's credentials (e.g., employment status, academic status), in response to a request by a requestor (e.g., potential customer, employer, business partner) are disclosed. A validation application, running on a first device under control of a requestor, receives a request to validate one or more credentials of a credential holder. Responsive to the request, the validation application generates a unique link for presenting to the credential holder, for submitting the one or more credentials of the credential holder. The validation application presents the unique link in a graphical user interface of the first device. The credential holder scans the unique link provided by the requestor. The unique link directs the credential holder to an interface for submitting validatable credentials to a mediator service. The first device receives, from the mediator service, a validation response associated with the one or more credentials submitted by the credential holder.
G06F 21/34 - Authentification de l’utilisateur impliquant l’utilisation de dispositifs externes supplémentaires, p. ex. clés électroniques ou cartes à puce intelligentes
G06K 7/14 - Méthodes ou dispositions pour la lecture de supports d'enregistrement par radiation électromagnétique, p. ex. lecture optiqueMéthodes ou dispositions pour la lecture de supports d'enregistrement par radiation corpusculaire utilisant la lumière sans sélection des longueurs d'onde, p. ex. lecture de la lumière blanche réfléchie
G06Q 50/26 - Services gouvernementaux ou services publics
30.
SYSTEM AND METHOD FOR USE WITH AN ENTERPRISE APPLICATION ENVIRONMENT FOR DETERMINATION OF PEER BENCHMARKS
In accordance with an embodiment, described herein is a system and method for use with an enterprise application environment for generating data analytics, including determination and visualization of peer benchmarks. In accordance with an embodiment, the system operates to receive an enterprise data, for example from an organization's enterprise software environment, and provides computer-executable processes and data structures for use in configuring and generating peer performance indicators or benchmarks. In accordance with an embodiment, the system provides a user interface by which a user can define templates, domains, segments, and members that comprise peer groups, or customize peer definitions to suit particular requirements. The system can then assess the enterprise data across various dimensions or measures, and generate key performance indicators, dashboards, or scorecards, for example as a two-dimensional data visualization that compares peer-to-peer characteristics, benchmarks, or other insights.
Duality views are object views that return JSON duality view objects ("JDV objects"). JDV objects are virtual because they are not stored as JSON objects. Rather, JDV objects are stored in normalized form across "base rows" of "base tables" and "base columns". Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways. A notification directive specifies which of these ways causes a change to the JDV object to be journaled.
G06F 16/28 - Bases de données caractérisées par leurs modèles, p. ex. des modèles relationnels ou objet
32.
PROVISIONING AND MANAGING RESOURCES WITHIN A CLOUD INFRASTRUCTURE OF A FIRST CLOUD SERVICE PROVIDER FOR A CLOUD SERVICE OFFERED BY A SECOND CLOUD SERVICE PROVIDER
Techniques are disclosed for provisioning and managing resources within a cloud infrastructure of a first cloud service provider for a cloud service offered by a second cloud service provider. Cross-cloud services can be provisioned and managed by and between private clouds of cloud service providers. The techniques include receiving a request for a cloud service by a component of a first private cloud within a first cloud environment and from a component of a second private cloud within a second cloud environment. The techniques further include the component of the first private cloud performing one or more operations to establish network connectivity prerequisites for network connectivity between the first private cloud and the second private cloud and causing one or more components of the first private cloud to provision the cloud service in the second private cloud.
H04L 67/10 - Protocoles dans lesquels une application est distribuée parmi les nœuds du réseau
G06F 9/455 - ÉmulationInterprétationSimulation de logiciel, p. ex. virtualisation ou émulation des moteurs d’exécution d’applications ou de systèmes d’exploitation
G06F 9/50 - Allocation de ressources, p. ex. de l'unité centrale de traitement [UCT]
H04L 41/0806 - Réglages de configuration pour la configuration initiale ou l’approvisionnement, p. ex. prêt à l’emploi [plug-and-play]
33.
Distributing Certificate Bundles According To Distribution Schedules
Operations of a certificate bundle distribution service may include: detecting a trigger condition to distribute a certificate bundle that includes a set of one or more certificate authority certificates; partitioning each particular network entity of a plurality of network entities associated with a computer network into one of a plurality of certificate distribution groups based on a network address of the particular network entity, in which each particular certificate distribution group includes a particular subset of network entities from the plurality of network entities; selecting a particular certificate distribution group, of the plurality of certificate distribution groups, for distribution of the certificate bundle; and transmitting the certificate bundle to the particular subset of network entities in the particular certificate distribution group.
A system synchronizes a server-side DOM tree and a browser-side DOM tree with one another. Server may receive from a browser, a hash value of the browser-side DOM tree, and a server-side update instruction for applying a first server-side update to the server-side DOM tree to synchronize with a first browser-side update by the browser to the browser-side DOM tree. The server may identify the server-side DOM tree based on the hash value. The server may execute upon the server-side DOM tree, the first server-side update and a second server-side update that is triggered by the first server-side update. The server may compute a browser-side update instruction for applying a second browser-side update to the browser-side DOM tree to synchronize with the server-side DOM tree. The server may transmit the browser-side update instruction to the browser, and the browser may apply the second browser-side update to the browser-side DOM tree.
G06F 16/958 - Organisation ou gestion de contenu de sites Web, p. ex. publication, conservation de pages ou liens automatiques
G06F 40/154 - Transformation en arborescence pour documents en configuration arborescente ou balisés, p. ex. langages XSLT, XSL-FO ou feuilles de style
G06F 40/194 - Calcul de la différence entre fichiers
Techniques for validating credentials are disclosed. A system receives a request, submitted by a requestor via a graphical user interface (GUI) of a first device under control of the requestor, to validate one or more credentials of a credential holder. The system transmits, to the first device under control of the requestor, a unique link to present in the GUI of the first device. The system receives, via the unique link from a second device under control of the credential holder, the one or more credentials of the credential holder. Responsive to receiving the one or more credentials of the credential holder, the system determines if the one or more credentials of the credential holder are valid and notifies the requestor of a result of determining if the one or more credentials are valid.
Innovative security is imposed on a database view in a relational database to ensure semantic integrity of database columns, some of which may be restricted columns that can be hidden, read only, referential, or value constrained. A computer hosts a database management system (DBMS) that generates a database view based on a data definition language (DDL) statement that specifies a value expression for a restricted column in the base table. A data manipulation language (DML) statement executes, which entails dynamic and innovative generation of values. Evaluating the value expression generates a new value for the base table. The new value is then stored into the restricted column in the base table. View initialization can be entirely confined to that single DDL statement. This approach does not use a database trigger and supports, but does not require, an ALTER DDL statement or a data control language (DCL) statement.
Changes made by a database transaction to one or more columns of a row in a database table are journaled to a change tracking table. Just as the column values as changed by a database transaction are visible to other database transactions when the database transaction commits, the changes journaled to the change tracking table are visible to other database transactions when the database transaction commits. A row in a change tracking table includes a column that contains the commit SCN of the changes recorded by the row. Commit patching provides transaction consistency between the commit SCN recorded for the row and the changes recorded by the row.
Duality views are object views that return JSON duality view objects ("JDV objects"). JDV objects are virtual because they are not stored as JSON objects. Rather, JDV objects are stored in normalized form across "base rows" of "base tables" and "base columns". Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. The changes made by a transaction are stored as a row in a history tracking table. When the transaction commits, its commit SCN is stored in associate with the row, even though the commit SCN is generated after the point in the transaction when the row itself can be modified.
Techniques for detecting and responding to network intrusions are disclosed. A network security system, executing in a computer network, accesses network details, including a network scope, a network definition, and/or an existing security policy associated with the computer network. The system further accesses network flow logs, including metadata that describes network traffic associated with the computer network. The system prompts a machine learning model to (a) identify anomalous network activity based at least in part on the network details and the network flow logs and (b) generate new network security policies when anomalous network activity is detected. The system receives, from the machine learning model, a new network security policy that addresses anomalous activity detected by the machine learning model. The system deploys the new security policy to the computer network without human review or approval of the new security policy.
Duality views are object views that return JSON duality view objects (“JDV objects”). JDV objects are virtual because they are not stored in a database as JSON objects. Rather, JDV objects are stored in normalized form across “base rows” of “base tables” and “base columns”. JDV objects are returned by a DBMS in response to database statements that request a JDV object from a JSON duality view; the JDV object that can be returned in this way may be referred to as belonging to or as being contained by and in the duality view. Importantly, through database statements that reference duality views, changes to the state of a JDV object may be specified at the level of a JDV object. A DBMS can execute a statement that assigns or sets a field value in a JDV object. The DBMS determines the particular underlying base tables, rows, and columns of the JDV object that need to be modified and then modifies those rows. The DBMS can execute a database statement that assigns a value to the base columns of those base rows. As a result, changes may be specified at the level of a JDV object (“object-directed change”) and may be specified to base columns of base rows of the JDV object (“table-directed change”). The ability to specify changes at these two levels may be referred to as data manipulation duality. The advancements to duality views described herein augment JSON objects with “computed” JSON fields, journal changes to JDV objects for replication and history tracking (“history tracking”) of changes, or validate changes made to JDV objects. The advancements are activated and controlled by, at least in part, issuing data definition language (“DDL”) commands that define directives that are associated with duality views. The computed fields of duality views are referred to as an augmented field. Augmentation directives for a duality view define one or more augmented fields for JDV objects belonging to the duality views and define how to compute field values of the augmented fields. The augmented fields and their respective field values are computed on the fly when, for example, generating JDV objects to return for database query statements that project JDV objects from duality views. Augmented fields may be computed from other JSON fields of a JDV object, including other augmented fields. Augmented fields may be generated by read augmentation or write augmentation. Augmented fields generated by read augmentation are similar to computed columns; these augmented fields are not stored in a base column of a base row. Write augmentation computes values for base columns, for example, default values. Notification directives define various ways of journaling changes to JDV objects. Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. The level of detail to journal includes the entire JDV object or just the changed portions. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways: by a table-directed change directed to the base row, an object-directed change to the JDV object, or an object-directed to change another JDV object in the same or different duality views. A notification directive specifies which of these circumstances causes a change to the JDV object to be journaled. A validation directive for a duality view specifies criteria that must be met by field values of the JDV object belonging to a duality view. The validation directives are applied to a JDV object in a duality view when the JDV object is changed or read. Similar to notification directives, validation directives specify what kind of changes trigger validation. A database transaction may change multiple JDV objects in multiple duality views. Notification directives and validation directives may require generating multiple JDV objects, the base rows for which may be numerous and may be stored in multiple tables; some of the base rows may be changed by the database transaction, and many may not. The multiple JDV objects to generate should be generated in a manner that is transactionally consistent. DBMS mechanisms for transactional consistency rely on row-level semantics that are insufficient for assuring transactional consistency at the object level.
Duality views are object views that return JSON duality view objects (“JDV objects”). JDV objects are virtual because they are not stored in a database as JSON objects. Rather, JDV objects are stored in normalized form across “base rows” of “base tables” and “base columns”. JDV objects are returned by a DBMS in response to database statements that request a JDV object from a JSON duality view; the JDV object that can be returned in this way may be referred to as belonging to or as being contained by and in the duality view. Importantly, through database statements that reference duality views, changes to the state of a JDV object may be specified at the level of a JDV object. A DBMS can execute a statement that assigns or sets a field value in a JDV object. The DBMS determines the particular underlying base tables, rows, and columns of the JDV object that need to be modified and then modifies those rows. The DBMS can execute a database statement that assigns a value to the base columns of those base rows. As a result, changes may be specified at the level of a JDV object (“object-directed change”) and may be specified to base columns of base rows of the JDV object (“table-directed change”). The ability to specify changes at these two levels may be referred to as data manipulation duality. The advancements to duality views described herein augment JSON objects with “computed” JSON fields, journal changes to JDV objects for replication and history tracking (“history tracking”) of changes, or validate changes made to JDV objects. The advancements are activated and controlled by, at least in part, issuing data definition language (“DDL”) commands that define directives that are associated with duality views. The computed fields of duality views are referred to as an augmented field. Augmentation directives for a duality view define one or more augmented fields for JDV objects belonging to the duality views and define how to compute field values of the augmented fields. The augmented fields and their respective field values are computed on the fly when, for example, generating JDV objects to return for database query statements that project JDV objects from duality views. Augmented fields may be computed from other JSON fields of a JDV object, including other augmented fields. Augmented fields may be generated by read augmentation or write augmentation. Augmented fields generated by read augmentation are similar to computed columns; these augmented fields are not stored in a base column of a base row. Write augmentation computes values for base columns, for example, default values. Notification directives define various ways of journaling changes to JDV objects. Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. The level of detail to journal includes the entire JDV object or just the changed portions. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways: by a table-directed change directed to the base row, an object-directed change to the JDV object, or an object-directed to change another JDV object in the same or different duality views. A notification directive specifies which of these circumstances causes a change to the JDV object to be journaled. A validation directive for a duality view specifies criteria that must be met by field values of the JDV object belonging to a duality view. The validation directives are applied to a JDV object in a duality view when the JDV object is changed or read. Similar to notification directives, validation directives specify what kind of changes trigger validation. A database transaction may change multiple JDV objects in multiple duality views. Notification directives and validation directives may require generating multiple JDV objects, the base rows for which may be numerous and may be stored in multiple tables; some of the base rows may be changed by the database transaction, and many may not. The multiple JDV objects to generate should be generated in a manner that is transactionally consistent. DBMS mechanisms for transactional consistency rely on row-level semantics that are insufficient for assuring transactional consistency at the object level.
Duality views are object views that return JSON duality view objects (“JDV objects”). JDV objects are virtual because they are not stored in a database as JSON objects. Rather, JDV objects are stored in normalized form across “base rows” of “base tables” and “base columns”. JDV objects are returned by a DBMS in response to database statements that request a JDV object from a JSON duality view; the JDV object that can be returned in this way may be referred to as belonging to or as being contained by and in the duality view. Importantly, through database statements that reference duality views, changes to the state of a JDV object may be specified at the level of a JDV object. A DBMS can execute a statement that assigns or sets a field value in a JDV object. The DBMS determines the particular underlying base tables, rows, and columns of the JDV object that need to be modified and then modifies those rows. The DBMS can execute a database statement that assigns a value to the base columns of those base rows. As a result, changes may be specified at the level of a JDV object (“object-directed change”) and may be specified to base columns of base rows of the JDV object (“table-directed change”). The ability to specify changes at these two levels may be referred to as data manipulation duality. The advancements to duality views described herein augment JSON objects with “computed” JSON fields, journal changes to JDV objects for replication and history tracking (“history tracking”) of changes, or validate changes made to JDV objects. The advancements are activated and controlled by, at least in part, issuing data definition language (“DDL”) commands that define directives that are associated with duality views. The computed fields of duality views are referred to as an augmented field. Augmentation directives for a duality view define one or more augmented fields for JDV objects belonging to the duality views and define how to compute field values of the augmented fields. The augmented fields and their respective field values are computed on the fly when, for example, generating JDV objects to return for database query statements that project JDV objects from duality views. Augmented fields may be computed from other JSON fields of a JDV object, including other augmented fields. Augmented fields may be generated by read augmentation or write augmentation. Augmented fields generated by read augmentation are similar to computed columns; these augmented fields are not stored in a base column of a base row. Write augmentation computes values for base columns, for example, default values. Notification directives define various ways of journaling changes to JDV objects. Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. The level of detail to journal includes the entire JDV object or just the changed portions. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways: by a table-directed change directed to the base row, an object-directed change to the JDV object, or an object-directed to change another JDV object in the same or different duality views. A notification directive specifies which of these circumstances causes a change to the JDV object to be journaled. A validation directive for a duality view specifies criteria that must be met by field values of the JDV object belonging to a duality view. The validation directives are applied to a JDV object in a duality view when the JDV object is changed or read. Similar to notification directives, validation directives specify what kind of changes trigger validation. A database transaction may change multiple JDV objects in multiple duality views. Notification directives and validation directives may require generating multiple JDV objects, the base rows for which may be numerous and may be stored in multiple tables; some of the base rows may be changed by the database transaction, and many may not. The multiple JDV objects to generate should be generated in a manner that is transactionally consistent. DBMS mechanisms for transactional consistency rely on row-level semantics that are insufficient for assuring transactional consistency at the object level.
A control plane of a cloud environment generates an application programming interface (API), that is configured to instantiate a plurality of network resources for a customer of the cloud environment. The API creates a first data object including first information identifying location of a rack within a network fabric, wherein the rack houses at least some of the plurality of network resources for the customer. Further, the API creates a second data object including second information corresponding to a cluster of network resources associated with the customer, wherein the cluster includes the plurality of network resources that are utilized to execute a workload for the customer. In response to receiving a request from the customer, the API instantiates the plurality of network resources across the network fabric utilizing the first data object and the second data object.
For a network of switches included in a multi-tier network fabric, an uplink port of each switch included in a first tier of switches is partitioned into a first plurality of sub-ports. A downlink port of each switch included in a second tier of switches is partitioned into the first plurality of sub-ports. For each uplink port of each switch included in the first tier of switches, a plurality of subsets of sub-ports is generated, where each subset of sub-ports of the switch is communicatively coupled to a different switch included in the second tier of switches. A packet is transmitted from a first GPU coupled to a first switch included in the first tier of switches to a second GPU coupled to a second switch included in the first tier of switches via utilizing one of the sub-ports of the uplink port of the first switch.
Techniques for updating a sentiment analysis model in response to feedback for operations that are executed based on an output of the sentiment analysis model are disclosed. The sentiment analysis model analyzes a chat conversation to determine user sentiment. The system executes an operation based on the user sentiment determined by the sentiment analysis model. The operation may be the transfer of the chat conversation from a chatbot to a human agent or the generation of an outbound message by the chatbot. The system receives positive or negative feedback regarding the appropriateness and/or timeliness of the operation. The positive or negative feedback is attributed to the sentiment analysis model. The system generates training data using the feedback. The system then retrains the sentiment analysis model based on the training data.
G06F 40/35 - Représentation du discours ou du dialogue
H04L 51/02 - Messagerie d'utilisateur à utilisateur dans des réseaux à commutation de paquets, transmise selon des protocoles de stockage et de retransmission ou en temps réel, p. ex. courriel en utilisant des réactions automatiques ou la délégation par l’utilisateur, p. ex. des réponses automatiques ou des messages générés par un agent conversationnel
46.
TECHNIQUES FOR TRANSFORMING NATURAL LANGUAGE CONVERSATION INTO A VISUALIZATION REPRESENTATION
Techniques are disclosed herein for transforming natural language conversations into a visual output. In one aspect, a computer-implement method includes generating an input string by concatenating a natural language utterance with a schema representation comprising a set of entities for visualization actions, generating, by a first encoder of a machine learning model, one or more embeddings of the input string, encoding, by a second encoder of the machine learning model, relations between elements in the schema representation and words in the natural language utterance based on the one or more embeddings, generating, by a grammar-based decoder of the machine learning model and based on the encoded relations and the one or more embeddings, an intermediate logical form that represents at least the query, the one or more visualization actions, or the combination thereof, and generating, based on the intermediate logical form, a command for a computing system.
Techniques may include receiving, by a cluster management system, a ruleset that defines a node operation and a condition to trigger the node operation. The techniques can further include receiving, by the cluster management system, a node identifier and a label for a node identified by the node identifier. The techniques can further include responsive to receiving the node identifier and the label, transmitting, by the cluster management system and to a service provider system, a request to perform an operation with respect to a containerized application that is executing on one or more computing devices, wherein the containerized application is one of a plurality of containerized applications that are executing on the one or more computing devices, and wherein the containerized application is identified by the node identifier.
An aspect of the present disclosure facilitates validating predicate-based filters used in asynchronous messaging systems. In one embodiment, a digital processing system identifies a predicate-based filter sought to be validated, and parses the predicate-based filter to determine a set of element identifiers contained in the predicate-based filter. The system generates a payload with each of the set of element identifiers assigned to a corresponding value and executes the predicate-based filter against the payload. The system indicates that the predicate-based filter is valid if a status of executing is a success, and is invalid otherwise.
Techniques for using conversational artificial intelligence to understand and manage source code repositories in software development environments. In one aspect, a computer-implemented method is described that includes receiving an utterance concerning code in a codebase, generating an utterance vector embedding based on the utterance, executing a similarity search on a vector index by comparing the utterance vector embedding to vector embeddings stored in the vector index using a similarity metric and identifying relevant nodes in a graph-based representation of the codebase based on the comparing, executing, based on types of constructs associated with the relevant nodes, queries on a graph database to retrieve additional context associated with the relevant nodes, generating a prompt comprising the utterance, the relevant nodes, and the additional context associated with the relevant nodes, and generating, by a generative model, a response to the utterance based on the prompt.
Systems, methods, and other embodiments associated with use of an LLM to validate infrastructure designs are described. In one embodiment, a method accesses a candidate topology for a compute infrastructure and infrastructure requirements that are in human language and are applicable to the candidate topology. The method generates a prompt from the candidate topology and infrastructure requirements, wherein the prompt is configured to cause a validation LLM to validate conformance of the candidate topology to the infrastructure requirements. The method generates, with the validation LLM, a status of validation for the candidate topology in response to the prompt. And, the method autonomously determines, based on the status of validation, whether to (1) issue a configuration instruction to configure a target computer system to have the compute infrastructure or (2) issue a correction instruction to generate a corrected topology when the status of validation is invalid.
H04L 41/16 - Dispositions pour la maintenance, l’administration ou la gestion des réseaux de commutation de données, p. ex. des réseaux de commutation de paquets en utilisant l'apprentissage automatique ou l'intelligence artificielle
51.
Sentiment And Alignment Scoring As A Proxy For Explicit User Feedback In Conversational Artificial Intelligence Systems
Artificial intelligence chatbot performance is improved through automated conversation quality analysis and training example generation. Chatbot conversations are processed via dual pathways, a sentiment analysis model measuring user satisfaction and an alignment analysis model evaluating query-answer semantic correspondence. These analyses produce alignment and sentiment scores that are combined into a sentiment and alignment deviation composite score called a “SADI score.” The SADI score quantifies response accuracy and satisfaction changes across conversation segments, enabling identification of problematic patterns. When SADI scores indicate significant deviations from performance thresholds, the new training examples are generated from the problematic conversations. The generated examples focus on segments showing sentiment or alignment degradation. A target machine learning model receives parameter updates using these examples, creating a feedback loop where user interactions drive model refinement. The systematic approach enables chatbot adaptation based on real interactions while addressing manual analysis limitations and implicit feedback incorporation.
A system receives a request from an application for a file system to execute a target operation corresponding to a target entity. In response to receiving the request to execute the target operation, the file system executes a logging operation that includes writing metadata to a log entry of a data structure stored in the one or more persistent storage media. The system determines that execution of the logging operation is complete and transmits, to the application, a response to the request that is based on completion of the logging operation. The logging operation is different than the target operation requested by the application via the request. After transmitting the response to the application, the file system completes execution of the target operation.
Embodiments back up application data for an application in a database for a tenant in a multi-tenant cloud based system. Embodiments, using a REST API, extract metadata of the database that corresponds to the tenant. Based on the metadata, embodiments identify one or more workspaces of the database that correspond to the tenant. For each of the workspaces, embodiments call REST API operations to collect the metadata corresponding to the workspaces and collect table data corresponding to the workspaces. Embodiments upload the collected metadata and table data to a backup archive.
G06F 11/14 - Détection ou correction d'erreur dans les données par redondance dans les opérations, p. ex. en utilisant différentes séquences d'opérations aboutissant au même résultat
54.
ALLOWING UPDATABLE VIEWS TO SPECIFY COLUMN VALUES TO BE USED FOR UPDATE OR INSERT WITHOUT REQUIRING THE COLUMN TO BE VISIBLE TO THE VIEW
Innovative security is imposed on a database view in a relational database to ensure semantic integrity of database columns, some of which may be restricted columns that can be hidden, read only, referential, or value constrained. A computer hosts a database management system (DBMS) that generates a database view based on a data definition language (DDL) statement that specifies a value expression for a restricted column in the base table. A data manipulation language (DML) statement executes, which entails dynamic and innovative generation of values. Evaluating the value expression generates a new value for the base table. The new value is then stored into the restricted column in the base table. View initialization can be entirely confined to that single DDL statement. This approach does not use a database trigger and supports, but does not require, an ALTER DDL statement or a data control language (DCL) statement.
Systems, methods, and other embodiments associated with adaptive planning using LLMs are described. In one embodiment, a method includes receiving a digital representation of a question expressed in natural language for resolution by a large language model. The method includes acquiring digital information that is relevant to answering the question and absent from a digital context associated with the question. The method includes determining that a gap exists in the acquired information based on assessment by the large language model of whether the acquired information is sufficient to answer the question. The method generates a sub-question based on detecting divergent dimensions among the embeddings, wherein the sub-question is formulated to request information that fills the gap in the acquired information. And the method includes recursively executing the computer-implemented method with the sub-question in place of the question until the gap in the acquired information is eliminated.
A control plane of a cloud service provider (CSP) deploys a plurality of controllers, wherein each controller is configured to communicatively couple a substrate network of the CSP to a network fabric. Each controller of the plurality of controllers includes a plurality of network interface cards. The plurality of controllers are coupled in a serial and cyclic manner by: connecting a first port of a first network interface card (NIC) of each controller to a second port of a second NIC of another controller, connecting a second port of the first NIC of each controller to the substrate network of the CSP, and connecting at least one other NIC of each controller to the network fabric.
For a network of switches included in a multi-tier network fabric, an uplink port of each switch included in a first tier of switches is partitioned into a first plurality of sub-ports. A downlink port of each switch included in a second tier of switches is partitioned into the first plurality of sub-ports. For each uplink port of each switch included in the first tier of switches, a plurality of subsets of sub-ports is generated, where each subset of sub-ports of the switch is communicatively coupled to a different switch included in the second tier of switches. A packet is transmitted from a first GPU coupled to a first switch included in the first tier of switches to a second GPU coupled to a second switch included in the first tier of switches via utilizing one of the sub-ports of the uplink port of the first switch.
A control plane of a cloud environment generates an application programming interface (API), that is configured to instantiate a plurality of network resources for a customer of the cloud environment. The API creates a first data object including first information identifying location of a rack within a network fabric, wherein the rack houses at least some of the plurality of network resources for the customer. Further, the API creates a second data object including second information corresponding to a cluster of network resources associated with the customer, wherein the cluster includes the plurality of network resources that are utilized to execute a workload for the customer. In response to receiving a request from the customer, the API instantiates the plurality of network resources across the network fabric utilizing the first data object and the second data object.
G06F 9/50 - Allocation de ressources, p. ex. de l'unité centrale de traitement [UCT]
H04L 41/12 - Découverte ou gestion des topologies de réseau
H04L 41/50 - Gestion des services réseau, p. ex. en assurant une bonne réalisation du service conformément aux accords
H04L 41/08 - Gestion de la configuration des réseaux ou des éléments de réseau
H04L 41/0895 - Configuration de réseaux ou d’éléments virtualisés, p. ex. fonction réseau virtualisée ou des éléments du protocole OpenFlow
H04L 41/04 - Architectures ou dispositions de gestion de réseau
H04L 41/0266 - Échange ou transport d’informations de gestion de réseau en utilisant l’InternetIntégration de serveurs de gestion du Web dans des éléments de réseauProtocoles basés sur les services du Web en utilisant des métadonnées, des objets ou des commandes pour formater l’information de gestion, p. ex. en utilisant un langage de balisage eXtensible [XML]
H04L 41/0273 - Échange ou transport d’informations de gestion de réseau en utilisant l’InternetIntégration de serveurs de gestion du Web dans des éléments de réseauProtocoles basés sur les services du Web en utilisant des services du Web pour la gestion du réseau, p. ex. le protocole d’accès aux objets simples [SOAP]
H04L 41/0806 - Réglages de configuration pour la configuration initiale ou l’approvisionnement, p. ex. prêt à l’emploi [plug-and-play]
H04L 41/0893 - Affectation de groupes logiques aux éléments de réseau
H04L 41/40 - Dispositions pour la maintenance, l’administration ou la gestion des réseaux de commutation de données, p. ex. des réseaux de commutation de paquets en utilisant la virtualisation des fonctions réseau ou ressources, p. ex. entités SDN ou NFV
59.
Upgrade Management in a Resource-Constrained Environment
Techniques are disclosed for upgrading an old version of an executable load file deployed on a secure element hardware platform into a new version of the executable load file while retaining static resources from the old version of the executable load file in the memory of the secure element hardware platform. The system identifies static resources from the old version of the executable load file to save in memory based on (a) instructions in an upgrade command that control the upgrade process and/or (b) requests by application instances derived from the old version of the executable load file. Unsaved resources from the old version of the executable load file are discarded, and the new version of the executable load file is deployed onto the secure element hardware platform. The static resources of the old executable load file are then selectively restored to the new version of the executable load file.
Duality views are object views that return JSON duality view objects (“JDV objects”). JDV objects are virtual because they are not stored in a database as JSON objects. Rather, JDV objects are stored in normalized form across “base rows” of “base tables” and “base columns”. JDV objects are returned by a DBMS in response to database statements that request a JDV object from a JSON duality view; the JDV object that can be returned in this way may be referred to as belonging to or as being contained by and in the duality view. Importantly, through database statements that reference duality views, changes to the state of a JDV object may be specified at the level of a JDV object. A DBMS can execute a statement that assigns or sets a field value in a JDV object. The DBMS determines the particular underlying base tables, rows, and columns of the JDV object that need to be modified and then modifies those rows. The DBMS can execute a database statement that assigns a value to the base columns of those base rows. As a result, changes may be specified at the level of a JDV object (“object-directed change”) and may be specified to base columns of base rows of the JDV object (“table-directed change”). The ability to specify changes at these two levels may be referred to as data manipulation duality. The advancements to duality views described herein augment JSON objects with “computed” JSON fields, journal changes to JDV objects for replication and history tracking (“history tracking”) of changes, or validate changes made to JDV objects. The advancements are activated and controlled by, at least in part, issuing data definition language (“DDL”) commands that define directives that are associated with duality views. The computed fields of duality views are referred to as an augmented field. Augmentation directives for a duality view define one or more augmented fields for JDV objects belonging to the duality views and define how to compute field values of the augmented fields. The augmented fields and their respective field values are computed on the fly when, for example, generating JDV objects to return for database query statements that project JDV objects from duality views. Augmented fields may be computed from other JSON fields of a JDV object, including other augmented fields. Augmented fields may be generated by read augmentation or write augmentation. Augmented fields generated by read augmentation are similar to computed columns; these augmented fields are not stored in a base column of a base row. Write augmentation computes values for base columns, for example, default values. Notification directives define various ways of journaling changes to JDV objects. Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. The level of detail to journal includes the entire JDV object or just the changed portions. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways: by a table-directed change directed to the base row, an object-directed change to the JDV object, or an object-directed to change another JDV object in the same or different duality views. A notification directive specifies which of these circumstances causes a change to the JDV object to be journaled. A validation directive for a duality view specifies criteria that must be met by field values of the JDV object belonging to a duality view. The validation directives are applied to a JDV object in a duality view when the JDV object is changed or read. Similar to notification directives, validation directives specify what kind of changes trigger validation. A database transaction may change multiple JDV objects in multiple duality views. Notification directives and validation directives may require generating multiple JDV objects, the base rows for which may be numerous and may be stored in multiple tables; some of the base rows may be changed by the database transaction, and many may not. The multiple JDV objects to generate should be generated in a manner that is transactionally consistent. DBMS mechanisms for transactional consistency rely on row-level semantics that are insufficient for assuring transactional consistency at the object level.
Duality views are object views that return JSON duality view objects (“JDV objects”). JDV objects are virtual because they are not stored in a database as JSON objects. Rather, JDV objects are stored in normalized form across “base rows” of “base tables” and “base columns”. JDV objects are returned by a DBMS in response to database statements that request a JDV object from a JSON duality view; the JDV object that can be returned in this way may be referred to as belonging to or as being contained by and in the duality view. Importantly, through database statements that reference duality views, changes to the state of a JDV object may be specified at the level of a JDV object. A DBMS can execute a statement that assigns or sets a field value in a JDV object. The DBMS determines the particular underlying base tables, rows, and columns of the JDV object that need to be modified and then modifies those rows. The DBMS can execute a database statement that assigns a value to the base columns of those base rows. As a result, changes may be specified at the level of a JDV object (“object-directed change”) and may be specified to base columns of base rows of the JDV object (“table-directed change”). The ability to specify changes at these two levels may be referred to as data manipulation duality. The advancements to duality views described herein augment JSON objects with “computed” JSON fields, journal changes to JDV objects for replication and history tracking (“history tracking”) of changes, or validate changes made to JDV objects. The advancements are activated and controlled by, at least in part, issuing data definition language (“DDL”) commands that define directives that are associated with duality views. The computed fields of duality views are referred to as an augmented field. Augmentation directives for a duality view define one or more augmented fields for JDV objects belonging to the duality views and define how to compute field values of the augmented fields. The augmented fields and their respective field values are computed on the fly when, for example, generating JDV objects to return for database query statements that project JDV objects from duality views. Augmented fields may be computed from other JSON fields of a JDV object, including other augmented fields. Augmented fields may be generated by read augmentation or write augmentation. Augmented fields generated by read augmentation are similar to computed columns; these augmented fields are not stored in a base column of a base row. Write augmentation computes values for base columns, for example, default values. Notification directives define various ways of journaling changes to JDV objects. Notification directives may define the replication of changes to JDV objects or may specify tracking changes in a form that can be queried. Notification directives specify what level of detail to journal and what kind of changes trigger journaling. The level of detail to journal includes the entire JDV object or just the changed portions. A JDV object of a duality view changes when a base row changes, which may also be a base row for another JDV object belonging to the same duality view or a different duality view. In addition, the base row may have been changed in multiple ways: by a table-directed change directed to the base row, an object-directed change to the JDV object, or an object-directed to change another JDV object in the same or different duality views. A notification directive specifies which of these circumstances causes a change to the JDV object to be journaled. A validation directive for a duality view specifies criteria that must be met by field values of the JDV object belonging to a duality view. The validation directives are applied to a JDV object in a duality view when the JDV object is changed or read. Similar to notification directives, validation directives specify what kind of changes trigger validation. A database transaction may change multiple JDV objects in multiple duality views. Notification directives and validation directives may require generating multiple JDV objects, the base rows for which may be numerous and may be stored in multiple tables; some of the base rows may be changed by the database transaction, and many may not. The multiple JDV objects to generate should be generated in a manner that is transactionally consistent. DBMS mechanisms for transactional consistency rely on row-level semantics that are insufficient for assuring transactional consistency at the object level.
In accordance with an embodiment, described herein is a system and method for use with an enterprise application environment for generating data analytics, including determination and visualization of peer benchmarks. In accordance with an embodiment, the system operates to receive an enterprise data, for example from an organization's enterprise software environment, and provides computer-executable processes and data structures for use in configuring and generating peer performance indicators or benchmarks. In accordance with an embodiment, the system provides a user interface by which a user can define templates, domains, segments, and members that comprise peer groups, or customize peer definitions to suit particular requirements. The system can then assess the enterprise data across various dimensions or measures, and generate key performance indicators, dashboards, or scorecards, for example as a two-dimensional data visualization that compares peer-to-peer characteristics, benchmarks, or other insights.
A network fabric including a first plurality of switches and a plurality of blocks of switches is provided where the first plurality of switches couples a plurality of host machines to the plurality of blocks of switches. Each block of switches includes a second plurality of switches arranged in a hierarchical manner including a first layer of switches that are coupled to a second layer of switches. A routing policy is configured for switches included in the first layer of switches, whereby the routing policy causes traffic received on a first port included in a first set of ports of the switch to be always forwarded on a second port included in a second set of ports of the switch. Upon a first GPU transmitting a packet destined for a second GPU, the packet is routed via the network fabric based on the configuring of the first layer of switches.
H04L 49/65 - Reconfiguration des commutateurs de paquets rapides
H04L 45/02 - Mise à jour ou découverte de topologie
H04L 49/253 - Routage ou recherche de route dans une matrice de commutation en utilisant l'établissement ou la libération de connexions entre les ports
64.
HIGH AVAILABILITY DESIGN FOR NETWORK FABRIC CONTROLLERS
A control plane of a cloud service provider (CSP) deploys a plurality of controllers, wherein each controller is configured to communicatively couple a substrate network of the CSP to a network fabric. Each controller of the plurality of controllers includes a plurality of network interface cards. The plurality of controllers are coupled in a serial and cyclic manner by: connecting a first port of a first network interface card (NIC) of each controller to a second port of a second NIC of another controller, connecting a second port of the first NIC of each controller to the substrate network of the CSP, and connecting at least one other NIC of each controller to the network fabric.
A network automation platform obtains first information corresponding to a network topology of a network fabric. The network automation platform retrieves second information corresponding to inventory information of the network fabric. The network automation platform generates a topology specification file based on merging the first information and the second information. Further, the network automation platform transmits the topology specification file to a controller, where the controller is configured to execute the network fabric based on the topology specification file.
Techniques are provided for generating context-aware, multi-turn natural language to SQL training data for database query models. The method involves accessing reference data that includes natural language utterances and corresponding queries in a programming query language. An AI model may generate a set of follow-up natural language utterances based on contextual information related to an initial utterance, and may rewrite these utterances to produce self-contained, multi-turn queries. An AI model can retrieve examples from the reference data that are semantically similar to the multi-turn utterances, serving as in-context examples. An AI model may then generate candidate queries based on the rewritten utterances and the retrieved examples. A dual stage validation process can be executed to evaluate the candidate queries, generating performance reports indicating which queries pass validation. Multi-turn training data comprising the validated utterances and their corresponding queries may then be generated for use in training database query models.
Techniques are disclosed herein to evaluating generative models for generating logical form queries. The techniques may include obtaining a query, wherein the query includes a natural language utterance; generating, by a generative model, a logical form query that corresponds to the natural language utterance; generating a normalized logical form query by modifying the logical form query based on a schema that defines how data is stored; executing the normalized logical form query on a data store to obtain results; identifying whether the results include at least reference results associated with execution of a reference logical form query on the data store; and indicating, in response to identifying that the results include at least the reference results, that the logical form query is acceptable.
Techniques are provided for ambiguity detection in and context-driven optimization in text-to-SQL translation. The method involves receiving a natural language utterance and a database context. A generative model may determine that ambiguity exists in the utterance, the database context, or both, by applying a context adherence factor that balances schema-based logical reasoning with common-sense interpretation. The generative model can further determine that the ambiguity corresponds to one of several predefined ambiguity categories and may classify the ambiguity by type based on the identified category. The system can generate an explanation describing the ambiguity and may formulate one or more reactive questions requesting information needed to resolve the ambiguity. An ambiguity score may be generated to indicate a level of confidence that the ambiguity is present. The system can provide to a user the ambiguity type, the explanation, the reactive questions, and the ambiguity score.
G06F 40/58 - Utilisation de traduction automatisée, p. ex. pour recherches multilingues, pour fournir aux dispositifs clients une traduction effectuée par le serveur ou pour la traduction en temps réel
Techniques include receiving, from a data store, a first request that comprises a node identifier identifying a node and indicating that a configuration of the node is to be updated to an updated configuration. The techniques include determining whether the configuration is different from the updated configuration. In accordance with determining the configuration is different: requesting the node be cordoned, requesting the node be drained, requesting to replace a boot volume of the node, associating the updated configuration with the node, and requesting the node be uncordoned. The techniques further including transmitting a status indicator to the data store to indicate that the first request was successful
G06F 11/14 - Détection ou correction d'erreur dans les données par redondance dans les opérations, p. ex. en utilisant différentes séquences d'opérations aboutissant au même résultat
G06F 11/07 - Réaction à l'apparition d'un défaut, p. ex. tolérance de certains défauts
G06F 11/16 - Détection ou correction d'erreur dans une donnée par redondance dans le matériel
Techniques for operating an upgrade-safe, extensible software-as-a-service (SaaS) platform are disclosed, including: receiving, by the SaaS platform, an application programming interface (API) request associated with a tenant of the SaaS platform; determining, by the SaaS platform, that the API request includes a tenant-specific type that extends a built-in type of the SaaS platform; responsive to determining that the API request includes the tenant-specific type: executing, by the SaaS platform, a mapper stack including at least (a) a first type-mapper associated with the built-in type of the SaaS platform and (b) a second type-mapper associated with the tenant-specific type; transmitting, by the SaaS platform, output of the mapper stack in response to the API request.
Techniques for presenting information indicating infrastructure security and compliance of services of a customer cloud environment to a customer-facing dashboard are disclosed. The information presented to the customer-facing dashboard is a subset of information available to operators associated with a cloud service provider (CSP). A tier-one dashboard service obtains information indicating infrastructure security and compliance of services in the customer cloud infrastructure environment. The tier-one dashboard service presents the information indicating infrastructure security and compliance to CSP operators on a CSP-facing dashboard. The CSP-facing dashboard is not accessible by customer operators. A tier-two dashboard service obtains the infrastructure security and compliance information from the tier-one dashboard service and filters the infrastructure security information to create a subset of information indicating the security and compliance of the services. The subset of infrastructure security and compliance information is presented to operators associated with a customer of the CSP on a customer-facing dashboard.
Techniques are disclosed herein for training and deploying a named entity recognition model. The techniques include implementing a nested labeling scheme for named entities within the training data and then training a machine learning model on the training data The techniques further include extracting an entity hierarchy for a predicted class based on a hierarchical template associated with a composite label, where the predicted class is representative of multiple named entity classes comprising at least a parent class and a child class associated with the composite label. The techniques further include increasing the volume of training data via data mining for sequence tags in a language corpus and then training a machine learning model on the training data.
Techniques for transmitting data within a cloud environment are disclosed. Data is received by a transmission service, which transmits the data to a signature and encryption (SE) service, along with an encryption key. Signed and encrypted data, which is received from the SE service, is (i) encrypted using the encryption key and (ii) signed by the SE service. In an example, the SE service maintains a log of the data. The signed and encrypted data is transmitted to an intermediate zone, to facilitate the intermediate zone to verify a signature of the signed and encrypted data, and allow passage of the signed and encrypted data to a reception service. The transmission and reception services are within a first tenancy and a second tenancy, respectively, of a cloud environment; and the intermediate zone is within one of the first tenancy or a third tenancy of the cloud environment.
H04L 9/32 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité comprenant des moyens pour vérifier l'identité ou l'autorisation d'un utilisateur du système
G06F 21/62 - Protection de l’accès à des données via une plate-forme, p. ex. par clés ou règles de contrôle de l’accès
G06F 21/64 - Protection de l’intégrité des données, p. ex. par sommes de contrôle, certificats ou signatures
Techniques are described for creating a container environment that implements direct communications between pods. This includes the creation of a management component (such as a control plane) of the container environment, the creation of nodes within an implementation component (such as a data plane) of the container environment, and the attaching of virtualized network interface cards (VNICs) to each node. This also includes creating a custom resource for each node, where the custom resource stores internet protocol (IP) addresses assigned to the node and stored within the VNICs for the node. Pods implemented within these nodes obtain an IP address from the VNICs for their respective node and communicate with each other utilizing the IP addresses via virtual communication devices and the VNICs.
H04L 41/08 - Gestion de la configuration des réseaux ou des éléments de réseau
G06F 9/455 - ÉmulationInterprétationSimulation de logiciel, p. ex. virtualisation ou émulation des moteurs d’exécution d’applications ou de systèmes d’exploitation
A change tracking and delta generation (CTDG) system servicing multiple clients is disclosed for enabling the delta generation during replication to be proportional to the number of changes between two snapshots instead of the size of the snapshots. In some embodiments, CTDG can track changes to a file system, where change-tracking can be dynamically switched between two sub-CT modes, file level or directory level. Change-tracking (CT) keys may be generated accordingly and stored in a data structure to be used for delta generation during replications for clients. In some embodiments, change-tracking may be enabled or disabled for different clients. CTDG can coordinate the CT enablement/disablement to ensure replications can use the information in CT keys properly.
Techniques are described herein for generating, encoding, and processing blockchain tokens as real estate in an augmented or virtual reality ecosystem. In some embodiments, the techniques include an encoding schema that encodes physical location coordinates to the unique identifier of a blockchain token. For example, the location coordinates may include the latitude and longitude of a physical space on earth. Blockchain tokens that have unique token identifiers represented by encoded geospatial information are referred to herein as location-encoded blockchain tokens. By encoding the physical location coordinates of a geographic location, the token owner or renter may populate the geographic space with curated content.
G06Q 20/36 - Architectures, schémas ou protocoles de paiement caractérisés par l'emploi de dispositifs spécifiques utilisant des portefeuilles électroniques ou coffres-forts électroniques
G06Q 20/38 - Protocoles de paiementArchitectures, schémas ou protocoles de paiement leurs détails
77.
GRAPH-BASED AND MACHINE-LEARNING TECHNIQUES FOR EVENT CLUSTERING AND THREAT DETECTION
Techniques discussed herein relate to a cloud-based threat detection framework that processes detected events using a hybrid graph-based and machine-learning approach. A computing system receives a feature vector associated with a detected event, including a strong entity indicator and a weak entity indicator. The strong entity indicator is processed by a graph model to generate a graph label, which, along with the weak entity indicator, is processed by a machine-learning clustering model to obtain a cluster label for the event. A risk score is determined based on the cluster label, quantifying the event's severity and likelihood of being a threat. A notification message is generated based on the risk score and sent to a user interface for further action. This cloud-based service enables dynamic correlation of events, reduces false positives, and provides actionable insights for efficient threat detection in real-time environments.
G06F 18/2321 - Techniques non hiérarchiques en utilisant les statistiques ou l'optimisation des fonctions, p. ex. modélisation des fonctions de densité de probabilité
78.
EFFICIENT STORAGE ALLOCATION BASED ON SUPPLY AND DEMAND MATCHING
A computer program product, system, and computer implemented method for efficient storage allocation based on supply and demand matching is provide that continually adjusts the available demand characteristics to historic demand characteristics. Generally, the approaches provided herein provide new procedures to quantify the distribution of storage allocations. These storage allocations maybe thought of as chunks which may be different sizes and which are quantified for both available storage chunks (supply) and storage allocation requests (demand). Specifically, a dynamic allocation approach is used to continually adjusts the available allocations to match past demand as a proxy for future demand and to provide for continual adjustment of available supply by at least making allocation determinations that address deviations of the actual supply from the measured demand. The approaches provided herein are applicable to at least caches, memory, flash, and disks (logical or physical or both).
Techniques may include receiving from a tenant client, a request to perform an operation with respect to a containerized application that is executing on one or more computing devices wherein the request comprises a cluster identifier and a node identifier that identifies a node, wherein the containerized application is one of a plurality of containerized applications that are executing on the one or more computing devices, and wherein the containerized application is identified by the cluster identifier and the node identifier. The techniques can further include providing a request to protect a workload of the node to the one or more computing devices identified by the cluster identifier and the node identifier. The techniques can further include providing a command to the node to cause the node to be repaired.
G06F 11/20 - Détection ou correction d'erreur dans une donnée par redondance dans le matériel en utilisant un masquage actif du défaut, p. ex. en déconnectant les éléments défaillants ou en insérant des éléments de rechange
80.
Executing Operations Corresponding To Target Entities Associated With File Systems
A system receives a request from an application to execute a target operation that includes removing a target directory. In response to receiving the request to execute the target operation, the system executes a parent directory-update operation that includes identifying a parent directory-attribute associated with the target directory and updating the parent directory-attribute to identify a first parent directory that is not accessible by the application. The system determines that execution of the parent directory-update operation is complete and transmits, to the application, a response to the request that is based on completion of the parent directory-update operation. The parent directory-update operation is different than the target operation requested by the application via the request. After transmitting the response to the application, the system executes the target operation.
A system receives a request from an application to execute a target operation that includes storing a data element in relation to a data abstraction unit. The system determines that storing the data element in relation to the data abstraction unit would result in an aggregate size corresponding to the data abstraction unit exceeding an application-defined size assigned to the first data abstraction unit. Responsive to determining that storing the data element in relation to the data abstraction unit would result in the aggregate size exceeding the application-defined size, the system rejects the request.
G06F 16/16 - Opérations sur les fichiers ou les dossiers, p. ex. détails des interfaces utilisateur spécialement adaptées aux systèmes de fichiers
G06F 16/13 - Structures d’accès aux fichiers, p. ex. indices distribués
82.
METHODS, SYSTEMS, AND COMPUTER READABLE MEDIA FOR PROVIDING HIERARCHICAL SECURITY EDGE PROTECTION PROXY (SEPP) DEPLOYMENT FOR ROAMING HUB USING SERVICE COMMUNICATION PROXY (SCP)-BASED MODEL D ROUTING
A method for providing a hierarchical SEPP deployment for a roaming hub using SCP-based model D routing includes registering, by SEPP instances and with an NRF instance in a combined roaming hub, NRF, and SCP deployment including the SEPP instances, the NRF instance and an SCP instance, NF profiles of the SEPP instances. The NF profiles each include a list of identifier(s) of PLMN(s) reachable through a registering SEPP instance. The SEPP instances heart-beat with the NRF instance. The method further includes obtaining, by the SCP instance and from the NRF instance, NF profiles of the SEPP instances in the combined roaming hub, NRF, and SCP deployment. The method further includes using, by the SCP instance, the NF profiles of the SEPP instances, including the lists of identifiers of PLMNs reachable through the SEPP instances, to route SBI request messages between the SEPP instances according to 3GPP communication model D.
H04L 41/0823 - Réglages de configuration caractérisés par les objectifs d’un changement de paramètres, p. ex. l’optimisation de la configuration pour améliorer la fiabilité
H04L 67/564 - Amélioration de la commande des applications basée sur des données interceptées des applications
83.
METHODS, SYSTEMS, AND COMPUTER READABLE MEDIA FOR RE-USING MONITORING KEY IN QUOTA MANAGEMENT BY POLICY CONTROL FUNCTION (PCF)
A method for re-using a monitoring key in quota management by a policy control function (PCF) includes receiving, from a session management function (SMF), a first session establishment request message for establishing a first packet data unit (PDU) session involving a subscriber. The method further includes fetching, by the PCF and from a unified data repository (UDR), a subscriber profile having a first usage monitoring profile including a first limitId. The method further includes creating a mapping between a monitoring key and the first limitId. The method further includes communicating, to the SMF, the monitoring key. The method further includes re-using the monitoring key for a second PDU session, the second PDU session having a second usage monitoring profile different from the first usage monitoring profile, the second usage monitoring profile including a second limitId different from the first limitId.
H04W 8/18 - Traitement de données utilisateur ou abonné, p. ex. services faisant l'objet d'un abonnement, préférences utilisateur ou profils utilisateurTransfert de données utilisateur ou abonné
Data about a plurality of incidents of a particular type in a computer environment can be stored. Each set of feature values of a plurality of sets of feature values for each incident corresponds to a different time period of a plurality of time periods. For each set of feature values, a composite value can be generated based on the set of feature values and the composite value is added to a set of composite values. A machine-learned model can generate predicted sets of aggregated values based on the set of composite values and based on different time period lengths. An aggregated set of values can be generated based on the predicted sets of aggregated values. An operation can identify and address issues based on the aggregated set of values.
A network automation platform (710) obtains first information corresponding to a network topology of a network fabric (715). The network automation platform (710) retrieves second information corresponding to inventory information of the network fabric (720). The network automation platform generates a topology specification file (727) based on merging the first information and the second information. Further, the network automation platform transmits the topology specification file to a controller (730), where the controller is configured to execute the network fabric based on the topology specification file.
H04L 41/0853 - Récupération de la configuration du réseauSuivi de l’historique de configuration du réseau en recueillant activement des informations de configuration ou en sauvegardant les informations de configuration
H04L 41/122 - Découverte ou gestion des topologies de réseau des topologies virtualisées, p. ex. les réseaux définis par logiciel [SDN] ou la virtualisation de la fonction réseau [NFV]
H04L 41/40 - Dispositions pour la maintenance, l’administration ou la gestion des réseaux de commutation de données, p. ex. des réseaux de commutation de paquets en utilisant la virtualisation des fonctions réseau ou ressources, p. ex. entités SDN ou NFV
H04L 49/356 - Interrupteurs spécialement adaptés à des applications spécifiques pour les réseaux de stockage
86.
Generative AI-based algorithm for improved cache hit performance in text-to-SQL systems
Techniques are disclosed herein for optimizing text-to-SQL systems using generative models to identify structurally equivalent requests in a cache to improve natural language conversion to SQL queries. In one aspect, a method includes: receiving a data request from a user. An embedding model embeds the data request generating a request embedding vector that is used to search a cache comprising stored requests. A candidate request is identified that comprises a curated natural language utterance that is structurally equivalent to the natural language utterance in the data request. The data request and the candidate request are transmitted to a generative model, that determines if the two requests are structurally equivalent and generates an SQL instruction corresponding to the data request using the SQL instruction associated with the candidate request as a template. The SQL instruction, a query result, or both are provided to the user.
Techniques for configuring a machine learning (ML) pipeline are disclosed. A system generates a region comprehension index (RCI) value that quantifies, for a visual dataset supplied to a computer vision model (CVM), a bias toward global or local reasoning. The system determines a first performance metric at least by applying the CVM to a whole image included in the visual dataset and a second performance metric at least by applying the CVM to patches of the whole image. The system determines the RCI value as a function of at least the first performance metric and the second performance metric. If the RCI value satisfies a criterion for deploying the CVM to an ML pipeline, the system selectively deploys the CVM to the ML pipeline.
Techniques for configuring a machine learning (ML) pipeline are disclosed. One or more embodiments generate a patch context robustness index (PCRI) value that quantifies, for computer vision model (CVM), sensitivity of the CVM to contextual noise, at least by: determining a first performance metric at least by applying the CVM to a whole image; determining a second performance metric at least by applying the CVM to patches of the whole image; and determining the PCRI value as a function of at least the first performance metric and the second performance metric. One or more embodiments deploy the CVM to an ML pipeline based at least on the PCRI value.
G06V 10/776 - ValidationÉvaluation des performances
G06V 10/70 - Dispositions pour la reconnaissance ou la compréhension d’images ou de vidéos utilisant la reconnaissance de formes ou l’apprentissage automatique
G06V 10/94 - Architectures logicielles ou matérielles spécialement adaptées à la compréhension d’images ou de vidéos
89.
SYSTEM AND METHOD FOR DOMAIN SPECIFIC MONITORING FOR APPLICATIONS
Systems, methods, and other embodiments associated with generation of monitoring applications are described. In one embodiment, a method includes generating a domain metamodel from a domain-specific target application. The domain metamodel includes metatags and hints that annotate a source code of the target application to denote domain elements. The method includes specifying relevancies of the domain elements to monitoring of performance characteristics of the target application. The method includes generating, using a code generator, instrumented code for the target application. The instrumented code includes monitoring code configured to monitor behavior of the domain elements. The instrumented code is generated based on the metatags and hints from the domain metamodel. The method includes building, using a compiler and runtime, the monitoring code to produce an executable monitoring application that monitors the target application when deployed in conjunction with the target application.
Techniques for hybrid retrieval using a knowledge base (KB) system for responsive multi-turn conversations are disclosed. The system evaluates user queries by generating confidence scores from comparisons against a set of predefined intents. When a confidence score falls within a high-confidence range, the system produces a predefined intent-based response. When a confidence score falls within a low-confidence range, the system retrieves a relevant KB document and generates a response by applying a language model (LM) to the query and the retrieved document.
Techniques are provided for a clustering algorithm to harvest in-context examples from Text-to-SQL logs. The method includes receiving vector representations of natural language utterances and corresponding queries in a programming query language. The clustering algorithm may be reparametrized using a training dataset by sampling the vector representations, estimating local radius parameter values from a distribution of pairwise distances, and iteratively executing the clustering algorithm while incrementally relaxing density constraints to identify paraphrases and unique queries. Hyperparameter tuning may be performed by adjusting clustering parameters based on evaluation metrics to obtain a fine-tuned algorithm. The method can further augment examples with ground truth examples and execute the fine-tuned clustering algorithm on the augmented set to generate clustered data points representing groups of paraphrases and unique queries. These clustered results may then be used to fine-tune or train a generative model, improving Text-to-SQL system performance and adaptability.
G06F 18/2321 - Techniques non hiérarchiques en utilisant les statistiques ou l'optimisation des fonctions, p. ex. modélisation des fonctions de densité de probabilité
G06F 16/22 - IndexationStructures de données à cet effetStructures de stockage
G06F 18/2137 - Extraction de caractéristiques, p. ex. en transformant l'espace des caractéristiquesSynthétisationsMappages, p. ex. procédés de sous-espace basée sur des critères de préservation de la topologie, p. ex. positionnement multidimensionnel ou cartes auto-organisatrices
G06F 18/22 - Critères d'appariement, p. ex. mesures de proximité
92.
PROHIBITING INCONSISTENT NAMED ENTITY RECOGNITION TAG SEQUENCES
A computer obtains a trained conditional random field (CRF) model including a set of model parameters learned from training data and stored in a transition matrix. Tag sequences, inconsistent with the tag sequence logic, are identified for the tags within the transition matrix, setting, within the transition matrix, a cost associated with transitioning between the pair of tags to be equal to a predefined hyperparameter value that penalizes the transitioning between the inconsistent pair of tags. The CRF model receives a string of text including one or more named entities. The CRF model inputs the string of text into the CRF model having the cost associated with the transitioning between the pair of tags set equal to the predefined hyperparameter value. The CRF model classifies the words within the string of text into different classes which might include the one or more named entities.
A source network virtualization device (NVD) receives a data packet destined for a cloud service accessible via a virtual IP address in a virtual cloud network (VCN). The source NVD encapsulates the data packet with a first encapsulation header to generate a first encapsulated data packet, the first encapsulated data packet being forwarded to a dynamic routing gateway (DRG). The DRG encapsulates the first encapsulated data packet with a second encapsulation header to generate a second encapsulated data packet, the second encapsulated data packet being forwarded to destination NVD associated with the cloud service, whereas the destination NVD decapsulates the second encapsulated data packet to obtain the first encapsulated data packet. The destination NVD communicates the first encapsulated data packet to the cloud service, wherein a virtual network interface card (VNIC) associated with the cloud service decapsulates the first encapsulated data packet to obtain the data packet.
H04L 41/40 - Dispositions pour la maintenance, l’administration ou la gestion des réseaux de commutation de données, p. ex. des réseaux de commutation de paquets en utilisant la virtualisation des fonctions réseau ou ressources, p. ex. entités SDN ou NFV
Disclosure relates to training a model to score video clips based on input prompts and associated ground truth information. Training datapoints may include a training video clip, an input prompt, and ground truth information. The model may generate a score for each clip based at least in part on the input prompt. A loss calculation subsystem may compute a loss using the score and the ground truth information. A loss minimization subsystem may update model parameters to train the model. In some embodiments, similarity scores may be generated for frames or images of a training video clip and/or be adjusted using a Weibull-based methodology to produce an overall clip score. In some embodiments, training may use an N-pair contrastive loss such that a clip annotated with a central time may receive a highest score while boundary clips annotated with a start time and/or end time retain partial similarity.
G06F 16/738 - Présentation des résultats des requêtes
G06F 16/783 - Recherche de données caractérisée par l’utilisation de métadonnées, p. ex. de métadonnées ne provenant pas du contenu ou de métadonnées générées manuellement utilisant des métadonnées provenant automatiquement du contenu
95.
PERFORMANCE OPTIMIZATION FOR QUALITY-OF-SERVICE FLOWS IN 5G NETWORK SLICES
The technology disclosed herein enables optimization of session performance in a network slice of a fifth-generation (5G) network. In a particular example, a method includes retrieving network information that indicates whether service levels are being met for sessions within a slice of the network slices. The method further includes classifying at least a portion of the sessions as unachievable sessions for which the service levels cannot be met due to limitations caused by radio link quality. When the network information indicates that a given service level is not being met, the method includes determining that a given session associated with the given service level is not one of the unachievable sessions and reconfiguring resource usage for the slice to achieve the given service level for the given session.
An augmentation and feature caching subsystem is described for training AI/ML models. In one particular aspect, a method is provided that includes receiving data comprising training examples, one or more augmentation configuration hyperparameters and one or more feature extraction configuration hyperparameters; generating a first key based on one of the training examples and the one or more augmentation configuration hyperparameters; searching a first key-value storage based on the first key; obtaining one or more augmentations based on the search of the first key-value storage; applying the obtained one or more augmentations to the training examples to result in augmented training examples; generating a second key based on one of the augmented training examples and the one or more feature extraction configuration hyperparameters; searching a second key-value storage based on the second key; obtaining one or more features based on the search of the second key-value storage.
The technology disclosed herein streamlines the design and provision of communication services. In a particular example, a method includes receiving service design parameters for a communication service and identifying, based on the service design parameters, a root node in a service design catalog. The method further includes generating a service design by traversing the service design catalog from the root node through one or more levels of child nodes defined by the service design catalog until one or more leaf nodes are reached. During the traversal, at each level, a child node is selected from among multiple candidate child nodes based on at least one of the service design parameters. Each candidate child node corresponds to a different configuration option for the communication service. The method also includes configuring one or more resource-facing services and one or more underlying resources corresponding to nodes included in the service design.
H04L 41/5006 - Création ou négociation de contrats, de garanties ou de pénalités au niveau du service [SLA]
H04L 41/5019 - Pratiques de respect de l’accord du niveau de service
98.
METHODS, SYSTEMS, AND COMPUTER READABLE MEDIA FOR UNIQUELY IDENTIFYING SESSION CONTEXTS FOR HANDLING SPIRAL CALLS IN A SESSION INITIATION PROTOCOL (SIP) CALL STATEFUL PROXY
A method for identifying session contexts for handling spiral calls includes, at a SIP session manager, receiving, from a first node, a first SIP request message. The method further includes determining that the first SIP request message is a request for establishing a new session or subscription. The method further includes, in response to determining that the first SIP request message is a request for establishing a new session or subscription, generating a first unique identifier. The method further includes using the first unique identifier as or to generate a first session context identifier. The method further includes creating a first session context database record for the first session or subscription. The method further include inserting the first unique identifier in a Record-Route header. The method further includes adding the Record-Route header to a first outbound SIP request message and routing the first outbound SIP request message.
Techniques for detecting inter-tenancy exfiltration of data in a cloud environment are disclosed. A log that includes information associated with receipt of a service message at a gateway within a cloud environment is accessed. Based on the log, (i) originating information of the service message (such as identification of an originating tenancy of the service message) and (ii) target information of the service message (such as identification of a target tenancy of the service message) are determined. The originating information and the target information are compared. A mismatch between the originating information of the service message and the target information of the service message is detected. In response to the detected mismatch between the originating information of the service message and the target information of the service message, information indicative of the detected mismatch is caused to be presented at a user interface.
G06F 21/57 - Certification ou préservation de plates-formes informatiques fiables, p. ex. démarrages ou arrêts sécurisés, suivis de version, contrôles de logiciel système, mises à jour sécurisées ou évaluation de vulnérabilité
G06F 21/55 - Détection d’intrusion locale ou mise en œuvre de contre-mesures
Techniques for conducting membership inference attacks are disclosed. In an example, a plurality of target interactions of a target user with an item providing platform are monitored. A plurality of target recommendations for the target user is received from a recommendation system of the item providing platform. Using an attack classifier and based on (i) the plurality of target interactions and (ii) the plurality of target recommendations, an inference is made as to whether at least a subset of the plurality of target interactions and/or at least a subset of the plurality of target recommendations were used to train the recommendation system. The attack classifier is trained using training data associated with a plurality of autonomous users (such as autonomous sock puppets) interacting with the item providing platform. In an example, the item providing platform is one of a video providing platform, an audio providing platform, or a shopping platform.
