Capital One Patente

Systems and methods for improving component access for cloud service users. The systems and methods allow for users to access additional cloud components through data layer operations such as routing table updates. Notably, the data layer may be modified by the cloud service users themselves and thus avoid restriction in inactivity by the cloud service provider.

A method and related system may establish communication guardrails during multi-turn conversations. The method may include constructing a graph indicating a set of intents and a set of intent states by inputting a set of transcripts to a language model and determining a set of metrics of events by traversing a path through nodes of the graph based on intent states indicated by the graph.

A device may receive an input indicating a prompt for a language model. The device may determine, using the language model, an output responsive to the prompt. The device may determine a degree by which the output relies on memory of the language model. The device may output, responsive to the degree by which the output relies on memory indicating an over-reliance on memory, a modified output that differs from the output.

A method and related system may reduce hallucinatory text generation by using shuffled training data.

Described herein are systems, methods, and programming for determining a context source of a large language model. For example, the techniques leverage a token accessible large language model to extract token activations associated with answers generated by the model in response to prompts. The token activations can be passed to a trained classification model that learns patterns of token activations associated with context-based model outputs. The trained classification model can assign a label to the model-produced answer indicating whether the answer was generated by the large language model using the input context included by the prompt or model memory of the large language model.

Methods and systems are described herein for facilitating queryless information retrieval during a live conversation via a bifurcated model. For example, the system may receive, a first set of decrypted utterances spoken during the live conversation. The system may provide the first set of decrypted utterances as input to a bifurcated model comprising (i) a first bifurcated model portion associated with a dynamic utterance embedding space and (ii) a second bifurcated model portion associated with a static computer file embedding space. The system may then generate, via the bifurcated model, a first intermediate output from the first bifurcated model portion indicating a first utterance embedding and a second intermediate output from the second bifurcated model portion indicating a first computer file embedding. The system may then generate during the live conversation, a graphical representation of a computer file associated with the first computer file embedding.

Systems and methods for generating plain text descriptions by translating predictions based on speech signal or data processing from similar users. The system may retrieve a first embedding set, wherein the first embedding set comprises a plurality of user embeddings, wherein each user embedding of the plurality of user embeddings, correspond to respective user query of a plurality of user queries. The system may process the plurality of user embeddings with a first classifier model to generate a first subset of user embeddings corresponding to a first prediction of a prediction pairing and a second subset of user embeddings corresponding to a second prediction of the prediction pairing. The system may determine, using the first classifier model, that a first user query corresponds to the first prediction. The system may determine a first response to the first user query based on the prediction pairing.

Methods and systems are described herein for managing prompts used for executing and evaluating supervised fine tuning (SFT) processes. For example, a model identifier and a first description of a first computing task may be extracted from a first request to execute an SFT process to perform the first computing task. A candidate prompt template and associated metadata including a candidate prompt name may be generated, and the candidate prompt template may be uploaded to a prompt database with the associated metadata. A second request to execute an SFT process indicating the candidate prompt name may be received. The candidate prompt template may be used to execute the SFT process on a large language model associated with the second request, resulting in a trained first large language model. The trained first large language model may be evaluated using a test data set and the candidate prompt template.

A device may obtain a plurality of responses of a language model to a prompt, where each of the plurality of responses comprises a binary response portion and a reasoning portion. The device may determine an uncertainty quantification for the language model based on the plurality of responses. The uncertainty quantification may be based on variations across reasoning portions of the plurality of responses, or variations in binary option confidences in binary response portions of the plurality of responses. The device may output the binary response portion of one of the plurality of responses based on the uncertainty quantification indicating certainty in the plurality of responses, or an uncertainty indication based on the uncertainty quantification indicating uncertainty in the plurality of responses.

In some implementations, a system may obtain, based on a graphical user interface (GUI) that includes a first field identifying a plurality of transformation objects, first input that indicates selection of a transformation object of the plurality of transformation objects. The system may obtain, based on a set of one or more second fields of the GUI that are associated with identifying parameter information associated with the transformation object, second input that indicates one or more parameters associated with the transformation object. The system may generate a transformation data structure, to be processed in a transformation sequence, that indicates the transformation object and the one or more parameters associated with the transformation object.

Systems, methods, and apparatuses are described for validating compliance in disaster recovery exercises (DREs). In conducting a DRE, a method may collect network data and may enrich the collected network data with application data. The method may simulate the unavailability of a cloud network node, collect network data in the new configuration, compare the two collected network data sets, and determine applications that rely on the simulated unavailable cloud network node. The method may indicate which application was resilient to the simulated unavailable network cloud node based on its determination. By enriching, the method may create a mapping to trace each traffic packet back to a specific application and verify whether the specific application's performance complies with the DRE. Unlike the current methods that lead to time-consuming manual analyses that are prone to false positives, the method may validate compliance in an automated way in a near-real-time fashion.

A device may retrieve data representing a plurality of activity sequences associated with a plurality of channels. Each activity sequence may include a plurality of activity entries having data values for a plurality of fields. The plurality of activity entries may be associated with respective timestamps. The plurality of activity sequences may have respective temporal resolutions. The device may flatten the data values for the plurality of activity sequences, across the plurality of activity entries, the plurality of fields, and the plurality of channels, into a single time-ordered activity sequence. The device may generate activity data embeddings using the single time-ordered activity sequence, generate, using an encoder module, an encoder output based on the activity data embeddings and temporal information based on the respective timestamps, input the encoder output into a decoder module, and generate, using the decoder module, a decoder output based on the encoder output.

In some implementations, a device may receive information identifying a cloud platform associated with generating a data log. The device may identify an application programming interface (API) associated with the cloud platform. The device may establish, using the API associated with the cloud platform, a listener service associated with receiving the data log asynchronously to an execution of a particular type of function in a cloud environment associated with the cloud platform. The device may receive, based on establishing the listener service and based on the execution of the particular type of function, the data log. The device may forward, to a primary endpoint associated with a data log platform and a failover endpoint associated with a data storage platform, and based on receiving the data log, information associated with the data log.

A zero-trust gateway system is provided for enforcing a zero-trust principle for granting access to one or more single sign-on (SSO) service provider entities provided in a zero-trust network. The zero-trust gateway system includes a plurality of interfaces for communicating with the one or more SSO service provider entities, one or more SSO identity provider entities, and one or more data sources provided in the zero-trust network; and one or more processors configured to: detect an authentication event associated with an access attempt for an SSO service provider entity of the one or more SSO service provider entities, the access attempt being indicated by an SSO identity provider entity of the one or more SSO identity provider entities; and authenticate or limit the access attempt based on information associated with the access attempt, the information being obtained from the one or more data sources.

A method and related system for reducing computational cost when using a natural language model for adaptive forms includes transcribing an utterance produced into a text portion during a conversation associated with a first person. The method also includes determining a result indicating that an interval indicating when a field of the set of data entry fields was last updated satisfies a data entry interval threshold. The method also includes providing the text portion to a natural language model to predict a set of intents based on the result and presenting at least one element of the second set of UI elements on the UI screen based on the set of intents.

Aspects of this disclosure relate to systems and methods for preventing injection attacks to large language models (LLMs). Techniques implemented by these systems and methods can include obtaining a query from a user device that is configured to cause an LLM to generate an output outside of a predetermined scope. In response to obtaining the query, systems can execute one or more evaluation models that are configured to identify one or more compatibility issues that can arise if the queries are executed by an LLM. Where a query is identified as having one compatibility issue(s), the systems described herein can either update the query to address the issue. Additionally, or alternatively, the systems described herein can prevent the processing of the query (and processing in furtherance of the injection attack) by an LLM.

Described herein are systems, methods, and programming for determining network resource code updates in real time while also reducing network resource consumption used to determine performance impacts of the network resource code updates. Updates to code portions may be generated and evaluated in a secure virtual computing space for an event of a sequence of events. If the updates to the code portions satisfy processing conditions, other events from the sequence may be evaluated against the same updates. A retroactive analysis of a portfolio of authorized accounts, each including a respective sequence of events, may be analyzed to determine the impact of the network resource code updates to the portfolio.

A system for improving scalability of machine learning may identify a machine learning libraries that are to be used for an execution of a serverless machine learning process using a serverless service provided by a cloud computing platform; remove a plurality of portions of one or more first libraries, of the machine learning libraries, that are to be pre-loaded into the serverless service; select one or more second libraries, of the machine learning libraries, that are to be dynamically loaded while the serverless machine learning process is executing; determine smart sampling of data to be used for the serverless machine learning process; and cause the execution of the serverless machine learning process by uploading a package that includes the one or more first libraries to the serverless service, causing the one or more second libraries to be dynamically loaded, and using the portion of the data.

Exemplary embodiments may use a contactless card as a secondary form of authentication in a multi-factor authentication for a secure messaging service. The recipient party of a request to initiate a messaging service session (such as a server computing device) may be programmed to use the phone number of the originating device to look up records regarding an identity of a party and their associated phone number as a primary credential and then may require an authentication credential originating from the contactless card as a secondary credential for the initiating party. In some instances, the credential originating from the contactless card is a onetime password that is valid only for a period of time. The recipient party determines whether the onetime password is valid. If both credentials are valid, a secure messaging session may be initiated with the initiating party.

In some implementations, a remote scanning system may receive, from a developer system, a request for a software package from a package repository. The remote scanning system may provide at least a portion of the software package to a machine learning model in order to receive a security indicator associated with the software package. The remote scanning system may generate a message based on the security indicator. The remote scanning system may transmit, to the developer system and in response to the request, the message. The remote scanning system may receive, from the developer system, a confirmation that the software package is requested. The remote scanning system may forward the software package from the package repository and to the developer system.

In some implementations, a system may receive, from a client device, a client request for a dynamic user experience associated with a user interface. The system may obtain an experience configuration associated with the dynamic user experience, wherein the experience configuration indicates a data provider for the dynamic user experience. The system may transmit, to the data provider, a data request, wherein the data request is generated based on the experience configuration. The system may receive, from the data provider, the dynamic data based on transmitting the data request. The system may transmit, to the client device, one or more tile specifications indicating user interface information for a user interface tile associated with the dynamic user experience to cause the client device to display the user interface tile via the user interface.

Various examples generally directed to unlocking a locked contactless card for use as payment in card-present transactions are provided. The contactless card is a payment card associated with a payment account. An application executing on a mobile device may be operable to enable unlocking the locked contactless card and receive an indication that the contactless card associated with the mobile device is unlocked for use in card-present transactions. The application may present information related to permitted uses of the unlocked contactless card via a graphical user interface. The permitted uses of the unlocked contactless card may be modified via the graphical user interface.

In some implementations, a system may receive, from a client device, a request for a dynamic user experience associated with a user interface. The system may receive, from the client device and in association with receiving the request for the dynamic user experience, capability information indicating one or more client capabilities supported by the client device for the dynamic user interface tiles. The system may transmit, to the client device, a tile specification indicating user interface information for a user interface tile associated with the dynamic user experience, where the tile specification is in accordance with the capability information and one or more provider capabilities.

In some implementations, a system may receive, from a client device, a client request for a dynamic user experience associated with a user interface, wherein the client request indicates client data associated with the client device. The system may obtain, based on the client data, dynamic data associated with the dynamic user experience and the client device. The system may provide, to a tile utility component of the system, tile definition information, wherein the tile definition information includes the client data, an indication of the dynamic user experience, and the dynamic data. The system may obtain, from the tile utility component, the one or more tile specifications based on the tile definition information. The system may transmit, to the client device, the one or more tile specifications to cause the client device to display the dynamic user experience via the user interface.

Aspects discussed herein may relate to methods and techniques for scanning application programming interface requests and responses to more readily identify data issues. The system may aggregate one or more data requests and/or responses according to correlated data transactions between devices. The system may then analyze traffic associated with those requests to determine if the responses are consistent with data policies. If a response is out of line with such policies, a system for reporting and correction are described.

The present disclosure relates to systems and methods for using multicomputer data transferring to active contactless communication. In one example, such a system may include at least one memory storing instructions and at least one processor configured to execute the instructions to perform one or more operations, the operations including: receiving a transaction associated with a merchant on a list of merchants stored in the at least one memory; determining, based on the merchant, a remote server associated with the merchant; assembling a data packet including an identifier of the contactless card and an identifier of the transaction; transmitting the data packet to the determined server across one or more computer networks; receiving confirmation from the determined sever in response to the data packet; and transmitting the confirmation to a user device associated with the transaction.

Methods and systems disclosed herein describe tokenizing data to generate a secure token that is limited in scope (e.g., directed to a specific recipient) and limited in time (e.g., valid for only a specified period of time). A detokenization process may be employed to recover encrypted data of the secure token without the need for any relational database lookup processes, thereby reducing cost while maintaining robust protection against unintended recipients that attempt to recover the encrypted data.

Systems and methods are disclosed for responding to a data incident. One or more processors receive one or more indicators of a data leak occurring at one or more nodes of a network. One or more processors causes an identification, by a machine-learning model, of one or more compromised nodes within the network based on the one or more indicators of a data leak. One or more processors may receive from the machine-learning model the identification of the one or more compromised nodes. One or more processors modify access permissions at one or more identified compromised nodes based on a user permission schema or pre-determined access rules, in response to the data leak. One or more processors cause a generation of a notification regarding the data leak and the modifications of access permissions to one or more users associated with the network.

Various embodiments are generally directed to techniques for generating rules automatically navigate a website and perform updates of payment token information.

Example embodiments of systems and methods for data transmission and authorizing transactions between transmitting and receiving devices are provided. In an embodiment, a method includes generating a request for at least one authentication factor when a value of the transaction exceeds a predetermined value, wherein the at least one authentication factor includes a cryptogram, receiving an authentication signal, the authentication signal including the cryptogram, and authorizing a transaction when the authentication signal is received.

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …

Zusammenfassung wird geladen …