Abstract: There are provided systems and methods for load balancing for map application route selection and output. A user may utilize a device application to map or route between two or more endpoints, such as geo-locations entered or detected by the device. During calculation of a travel route between the endpoints, real-time data, user preferences, and requesting entities may provide criteria data that may cause determination of a particular travel route, where the travel route may be longer than a most efficient route but within a pre-defined variable time or distance allotment and match the criteria data. Use of the route may accrue a form of compensation for the user. The user may view an application interface displaying the route, which may further include one or more executable processes to cause recalculation of the route. Recalculation of the route may require the user to provide credits or compensation.

Abstract: There are provided systems and methods for a carbon neutral blockchain protocol for resolving carbon offsetter payments for cryptocurrency transactions. A service provider, such as an online transaction processor, may provide a green wallet and network protocol that provides carbon offsetting fees and payments to carbon offsetters in order to account for carbon emissions and usage when processing cryptocurrency transactions on a blockchain network protocol. A green wallet may be provided, which allows registration of carbon offsetters and use of green addresses to cause cryptocurrency to be processed in a green transaction. A user's digital wallet may sign an initial cryptocurrency transaction and provide the fee to the green wallet. The green wallet may then resolve the cryptocurrency transaction by further signing the transaction and transmitting the fee to a digital wallet of the carbon offsetter over another cryptocurrency processing network.

Abstract: Techniques are disclosed relating to receiving, by a computer system from a user, a request to authorize the user. The technique may further include receiving, by the computer system, a live video stream of the user, and requesting, by the computer system, the user to submit an identification document from a list of authorized identification documents. While the user submits the identification document, the technique may also include detecting liveliness of the user in the live video stream by using a gesture recognition operation. Additionally, the technique may include determining, using the submitted identification document and the detected liveliness, whether to proceed with the request to authorize the user.

Abstract: Techniques are disclosed relating to a computer system identifying usage of a plurality of individual instances of a common computation task by a plurality of users of a networked service. These individual instances of the common computation task may generate a respective data set. Techniques also include creating, by the computer system, a global process to perform the common computation task. Execution of the global process may include generation of a global data set that includes at least portions of the respective data sets. Additionally, techniques include modifying, by the computer system, respective accounts of a subset of the plurality of users to use the global process in place of using a respective instance of the common computation task, as well as providing, by the computer system, the global data set generated by the global process to the respective accounts of the subset of users.

Abstract: Techniques are disclosed relating to transforming, by a computer system in response to a request to develop an evaluation strategy, the request into a first query for an artificial intelligence (AI) tool. The computer system may retrieve historical information that is related to the first query. The computer system may submit the first query to the AI tool. In response to receiving evaluation strategy information, the computer system may transform the evaluation strategy information into a second query for the AI tool. The computer system may submit the second query to the AI tool. In response to receiving executable code, the computer system may modify the executable code for compatibility with a particular computer system. The computer system may test performance of the modified executable code, and indicate that the modified executable code satisfies one or more performance thresholds for a code release.

Abstract: Techniques are disclosed relating to a graphical user interface (GUI) that is operable to depict data lineage information in levels. In some embodiments, data lineage information may specify a directed graph that is indicative of a data lineage associated with a plurality of data elements. For example, in the data lineage information, the plurality of data elements may be represented as a corresponding plurality of nodes and, in the directed graph, the plurality of nodes may be connected by edges in a manner that is indicative of the data lineage relationships between the plurality of data elements. In various embodiments, the disclosed techniques may generate a data lineage GUI that, for a selected data element of the plurality of elements, is usable to navigate different levels of the data lineage in an upstream and downstream direction relative to a particular level of the selected data element.

Abstract: Mechanisms for efficiently transferring multiple unspent transaction output (UTXO) tokens in a blockchain network operating a UTXO-based token transaction model are disclosed herein. These methods allow for the use of less computer processing and network bandwidth resources in the transfer of blockchain items, particularly in blockchain schemes with lineage tracking mechanisms (e.g. as may allow for clawback of tokens). Some embodiments comprise generating a delegated account and transferring the UTXO tokens into the delegated account. The ownership of the delegated account can then be transferred to another use of the blockchain network in a single transaction, thereby avoiding the need to individually transfer the UTXO tokens and incur the memory and computational resource burden and costs associated with therewith.

Abstract: Image-based processing of sources of financial information for financial transactions is provided. A device is enabled to acquire one or more images including a representation of a source of financial information. The device can validate the source, independent of the financial information in a preliminary analysis. If the source is validated, one or more images are used to acquire the financial information. The financial information is then validated. If the financial information is valid, authorization of the financial transaction can be initiated.

Abstract: There are provided systems and methods for on-device data privacy operations to prevent data sharing at user locations. A service provider, such as a merchant location of a merchant and/or associated online transaction processor, may provide additional services for to users via user data that may be tracked or stored of the user. However, the user may not want to share certain data with the merchant or other backend processor for privacy concerns. Thus, on-device data privacy operations may be used to detect when a user is at a location that has a corresponding privacy setting to hide or abstract user data for the location. The privacy setting may designate data to prevent from sharing when the user uses their device with devices associated with the location. Abstracted data associated with recommendations or actions to provide the user may be generated and provided to a merchant for the location.

Abstract: Techniques are disclosed relating to generating user-specific time-to-live (TTL) values using machine learning. In various embodiments, a server system maintains a cache data store that is operable to store data for a plurality of users of a web service. In response to a cache miss for a request from a first one of the plurality of users, the server system may generate a user-specific TTL value for the first user. In various embodiments, generating the user-specific TTL value may include using a machine learning model to generate a predicted future access pattern for the first user that indicates a distribution of time periods during which the first user is expected to access the web service and, based on the predicted future access pattern, determining the user-specific TTL value for the first user.

Abstract: Methods and systems are presented for providing explainable artificial intelligence for a deep-learning model on a global level. Multiple surrogate models are generated based on characteristics of the deep-learning model, where each surrogate model is configured to mimic a behavior of the deep-learning model with respect to one of the output dimensions associated with the deep-learning model. Simulations are performed on the surrogate models. Based on the simulation results, importance scores are calculated for each input feature of the deep-learning model. An importance score calculated for an input feature represents an extent to which the input feature contributes to a corresponding one of the output dimensions associated with the deep-learning model. The importance scores may then be used to modify the characteristics of the deep-learning model or other downstream machine learning models.

Abstract: Methods and systems are presented for providing a knowledge bot configurable to interact with users across multiple domains. The knowledge bot includes at least a text-based search engine and a semantic-based search engine. Each of the search engine is configured to retrieve documents from a corpus of documents based on the user query. The user query is in a natural language format. The retrieved documents may be ranked according to how relevant the documents are to the user query. A subset of the documents is used as the search results based on the ranking. The search results from the search engine are combined with the user query to generate a prompt for an artificial intelligence model. Based on the prompt, a response in the natural language format is generated by the artificial intelligence model.

Abstract: Techniques are disclosed relating to a method that includes receiving, by a particular computer system included in an enterprise computer system, deployment instructions, from a user, for a particular version of a machine-learning model. One or more versions of the machine-learning model may be stored in a database. The particular computer system may select, based on the deployment instructions, a destination within the enterprise computer system for deploying the particular version. The selected destination may provide access to a particular data set. The particular computer system may schedule a deployment of the particular version from the database to the selected destination. The deployed version of the machine-learning model may operate on the particular data set. Performance data associated with operation of the deployed version of the machine-learning model is collected.

Abstract: Techniques are disclosed relating to improving the efficiency of geolocation queries via geospatial cells and query parameter caching. In various embodiments, a system receives a request from a user device to access geolocation information of entities and determine a first geographic location of the device. The system identifies that the first location is within a geospatial cell stored in a database cache and determines whether query parameters corresponding to the cell are stored in the cache. The system may retrieve the cached query parameters corresponding to the geospatial cell and execute a geolocation query on a database storing location information for different entities based on the cached query parameters. Results of the geolocation query may be different from query results of one or more other queries executed, based on the cached query parameters, for other user computing devices located within the given geospatial cell based on the cached query parameters.

Abstract: There are provided systems and methods for providing a merchant recommendation using multiple models. An example system may receive, at a product predictor, an input including a series of products and generate an output including a target product associated with the series of products. The system may generate, at a value evaluator, a value change associated with the target product based at least in part on the output, and generate, at the product predictor, a recommendation including the target product and the value change associated with the target product. The system may further generate, at a seasonal predictor, a seasonal prediction for the output in a future time frame based on the output, and generate, at the product predictor, an updated output including an updated target product based on the seasonal prediction for the output in the future time frame.

Abstract: There are provided systems and methods for streamlined and privacy protected data flows for entity onboarding with online data platforms. An online transaction processor or other service provider may provide computing services and platforms to entities including merchants for electronic transaction processing and other account services. To onboard entities with the transaction processor, the transaction processor may provide a merchant or user-specific experience and recommendations using a conversational AI service and chatbot. An AI engine may be trained to engage with users via chat dialogue, which may interact with users during onboarding and/or account lifecycle events based on the available services and products of the service provider. An intent of the entity may be classified by the AI engine and may be used to generate a recommendation including an action plan or the like of activities or steps for the entity to take.

Abstract: There is provided systems and method for generating sale transaction from voice data input by a user. A user device may receive voice data including a preference for purchasing an item. The user device may convert the voice data to the preferences and perform a search for a sales transaction corresponding to the preferences. The search may include parameters about the user, such as a location. The sales transaction may include purchase prices, times, locations, or other relevant data. A user may accept or decline the sales transaction with additional user data. If the user accepts the sales transaction, the sales transaction may be completed with a payment provider and a transaction history given to the user for later redemption of the item. If the user declines the sales transaction, further sale transactions with additional items may be present to the user.

Abstract: Computing systems and methods for facilitating mobile transactions are disclosed. An example method includes: detecting, via a hardware communication interface, a user check-in request based at least on a communication exchange between a user device and a signaling device at a location. The communication exchange corresponds to a proximity of the user device to the signaling device. The method further includes: determining a user identifier associated with the user device based at least on the user check-in request; and transmitting one or more promotional items to the user device based at least on the user identifier.

Abstract: Systems and methods for processing machine learning attributes are disclosed. An example method includes: identifying a user transaction associated with a set of transaction attributes and a first transaction status; selecting, based on a risk evaluation model, a first plurality of transaction attributes from the set of transaction attributes; modifying a first value of a first transaction attribute in the first plurality of transaction attributes to produce a first modified plurality of transaction attributes; determining, based on the risk evaluation model, that the first modified plurality of transaction attributes identify a second transaction status different from the first transaction status; and in response to the determining, identifying the first transaction attribute as a risk attribute associated with the user transaction.

Abstract: Methods and systems for providing a stateless application framework are presented. The stateless application framework is utilized by different applications for implementing different workflows. Each workflow may be associated with one or more state machine configurations representing the different states within the workflow. Upon receiving an indication of an event from an application, a stateless application module transmits a job request to a data processing engine based on a current state of the event. When a response is received from the data processing engine, the state application module determines whether the state of the event has been updated since transmitting the job request. If it is determined that the state has been updated, the stateless application module is configured to transmit another job request to the data processing engine based on the updated state of the event.