Abstract: A system may include processor(s), and memory in communication with the processor(s) and storing instructions configured to cause the system to correct ASR errors. The system may receive a transcription comprising transcribed word(s) and may determine whether the transcribed word(s) exceed associated predefined confidence level(s). Responsive to determining a transcribed word does not exceed a predefined confidence level, the system may generate a predicted word. The system may calculate a distance between numerical representations of the transcribed word and the predicted word and may determine whether the distance exceeds a predefined threshold. Responsive to determining the distance exceeds the predefined threshold, the system may determine whether at least one red flag word of a list of red flag words corresponds to a context of the transcription, and, responsive to making that determination, may classify the transcription as associated with a first category.

Abstract: In some implementations, a credit decision platform may receive a credit request from an applicant and obtain domestic historical data associated with the applicant from a credit bureau device. The credit decision platform may obtain access to an email account associated with the applicant based on determining that the domestic historical data associated with the applicant is insufficient to process the credit request. The credit decision platform may identify, using one or more machine learning models, a set of email messages included in the email account that are relevant to the credit request and may analyze content included in the set of email messages to generate non-domestic historical data associated with the applicant. The credit decision platform may generate a decision on the credit request based on an estimated creditworthiness of the applicant, which may be determined based on the non-domestic historical data.

Abstract: Logic may provide a lock for a credit card that may, e.g., prevent or attenuate fraudulent transactions. Logic may include one or more user interaction mechanisms to physically interact with a user to change a user-configurable circuit to create an unlock pattern with the user-configurable circuit to allow the payment instrument to authorize transactions and one or more lock patterns to prevent the payment instrument from authorizing transactions. The one or more user interaction mechanisms may comprise moveable conductors to connect or disconnect circuits of the user-configurable circuit to create the unlock pattern and the lock patterns. The logic may comprise a processor to process transactions and be coupled with memory. The logic may receive power from a point of sale (POS) terminal to process a transaction, apply the power to the user-configurable circuit, and determine an output of the user-configurable circuit to determine if the transaction succeeds or fails.

Abstract: In some implementations, a regression prediction platform may obtain one or more feature sets related to an impending code change, wherein the one or more feature sets may include one or more features related to historical code quality for a developer associated with the impending code change or a quality of a development session associated with the impending code change. The regression prediction platform may provide the one or more feature sets to a machine learning model trained to predict a risk associated with deploying the impending code change based on a probability that deploying the impending code change will cause breakage after deployment and/or a probability that the impending code change will cause support issues after deployment. The regression prediction platform may generate one or more recommended actions related to the impending code change based on the risk associated with deploying the impending code change.

Abstract: Systems, methods, and apparatuses for proactive generation of biometric signatures are described. Based on a first biometric signature matching a first authenticated signature, an indication of portions of the biometric input that will be different after a second time may be generated. A second authenticated signature based on the indication and the biometric input received at the first time may be generated. A second biometric signature based on the indication and the biometric input received after a second time may be generated. Based on the second biometric signature matching the second authenticated signature, a prompt requesting authorization to generate a third authenticated signature based on the biometric input received after the second time may be generated. Based on receiving authorization to generate a third authenticated signature, a third authenticated signature may be generated.

Abstract: This disclosure provides an apparatus for a thermoelectric (TE) contrast therapy device and related methods. The thermoelectric contrast therapy device includes a control node (610) and a plurality of thermoelectric nodes (605). The control node electrically and mechanically connected to the plurality of thermoelectric nodes, the control node configured to direct the heating and cooling functions as well as timing functions. Each of thermoelectric nodes includes a thermoelectric module (105) structured with a supply side (115) and a waste side (120), a phase change material (110), a reservoir (725), a thermal sink (320), a spreader (125) and a controller (1010). The thermoelectric contrast therapy device pre-charges (1105) the PCM contained in the reservoir; provides (910) opposite heating and cooling functions using the supply side and the waste side of the thermoelectric module; and directs, using the controller, the heating and cooling functions as well as timing functions.

Abstract: Systems, methods, and apparatuses are described herein for performing sentiment analysis on electronic communications relating to one or more image-based communications methods, such as emoji. Message data may be received. The message data may correspond to a message that is intended to be sent but has not yet been sent to an application. Using a first machine learning model, one or more subsets of the plurality of emoji may be determined. The one or more subsets of the plurality of emoji may comprise one or more different types and quantities of emoji, and may each correspond to the same or a different sentiment. Using a second machine learning model, one or more emojis may be selected from the one or more subsets. The one or more emojis selected may correspond to responses to the message.

Abstract: Aspects provided may allow for an entity to deactivate a virtual card number (VCN). In particular, aspects provide for deactivating VCNs based on a frequency or periodicity of an interaction. After deactivation, the entity may block future charges to the VCN. Further aspects provide for using incoming charge data sent to one or more previous users to train a machine learning model to determine the frequency or periodicity of the interaction. Additional aspects may also provide for including trial length when determining the deactivation date, providing VCN configuration options to a user, and notifying the user based on upcoming and current VCN status changes.

Abstract: Methods, systems, and apparatuses are described herein for improving the accuracy of authentication questions using transaction limitations provided by users. A request for access to an account associated with a user may be received from a user device. An account restrictions database may be queried to determine one or more transaction rules associated with the account. The one or more transaction rules may have been created by the user and indicate limitations on financial transactions that may be performed via the account. An authentication question may be generated that is associated with a violation of the one or more transaction rules. The authentication question may be provided to the user device, and a response to the authentication question may be received. Access to the account may be provided to the user device based on the response.

Abstract: Disclosed embodiments may include a system configured to perform digital image analysis. The system may receive transaction data and image data associated with a user. The system may identify, from the transaction data, first travel feature(s). The system may identify, from the image data via computer vision, second travel feature(s). The system may train a machine learning model (MLM) to generate trip recommendation(s) for the user based on the first travel feature(s) and the second travel feature(s). The system may determine, via the trained MLM, whether at least a first trip recommendation of the trip recommendation(s) exceeds a predetermined threshold indicating a likelihood the user will be interested in the first trip recommendation. Responsive to determining the first trip recommendation exceeds the predetermined threshold, the system may provide the first trip recommendation to the user.

Abstract: Disclosed are methods and systems for generating a virtual tag for an item in a physical environment. For instance, an image of the physical environment captured by and displayed on a user interface of a computing device may be provided as input to a trained machine learning model configured to identify the item including a first subset of item attributes. The item may be identified and information associated with the item, including a second subset of item attributes, may be received from a data store using the first subset. A spatial location of the item may be identified in the image. An icon may be generated for the item and rendered by the computing device for display on the user interface in association with the spatial location of the item in the image. The icon may be indicative of at least one item attribute from the first or second subset.

Abstract: In some implementations, a system may receive interaction data corresponding to an interaction associated with a particular user of a plurality of users. The system may determine, based on the interaction data, one or more interaction categories associated with the interaction. The system may store the interaction data as historical interaction data associated with historical interactions of the particular user. The system may determine, based on the historical interaction data, experience statuses associated with the particular user in at least a subset of the plurality of categories. The system may receive, from a user device of the particular user, review data indicating a review associated with a historical interaction of the particular user. The system may store the review as historical review data. The review may be associated with one or more experience statuses of the particular user in the one or more interaction categories associated with the historical interaction.

Abstract: Systems and methods for external account authentication are disclosed herein. They include receiving a call to pair the external account with a secure account, extracting external data from the external account, the external data corresponding to external account content, providing user activity data from the secure account as an input to an authentication machine learning model, providing the external data as an input to the authentication machine learning model, the authentication machine learning model configured to output a certainty level that the external account is associated with a user of the secure account based on the external data and the activity data, receiving the certainty level from the authentication machine learning model, determining that the certainty level meets a certainty threshold, and pairing the external account with the secure account based on determining that the certainty level meets the certainty threshold.

Abstract: Aspects described herein may allow for authenticating a user by generating a customized set of authentication questions based on patterns that are automatically detected and extracted from user data. The user data may include transaction data collected over a period of time. By automatically detecting user patterns that correspond to user behavior over a period of time, an authentication system may be able to generate information that is recognizable to an authentic user but difficult to guess or circumvent for any other user.

Abstract: A non-transitory computer-readable storage medium storing computer-readable program code executable by a processor to retrieve a customer data record including bill payment activity associated with a provider account for a customer of a financial institution; and define a geographical footprint for the customer. The geographical footprint may be defined by identifying a customer set including the customer from a plurality of customers of a provider of the provider account, having customer records indicating geographic proximity, and determining that a predetermined fraction of the customer set are associated with a common billing cycle for the provider. The computer-readable program code may be executable by the processor to send a bill payment alert for transmission to the customer via electronic communication or telephonic communication, when the bill payment activity meets a threshold condition, wherein the threshold condition is determined at least in part by the geographical footprint.

Abstract: Systems and methods of the present disclosure enable intelligent dynamic caching of data by accessing an activity history of historical electronic activity data entries associated with a user account, and utilizing a trained entity relevancy machine learning model to predict a degree of relevance of each entity associated with the historical electronic activity data entries in the activity history based at least in part on model parameters and activity attributes of each electronic activity data entry. A set of relevant entities are determined based at least in part on the degree of relevance of each entity. Pre-cached entities are identified based on pre-cached entity data records cached on the user device, and un-cached relevant entities from the set of relevant entities are identified based on the pre-cached entities. The cache on the user device is updated to cache the un-cached entity data records associated with the un-cached relevant entities.

Abstract: A method and system for segmenting a plurality of images. The method comprises the steps of segmenting the image through a novel clustering technique that is, generating a composite depth map including temporally stable segments of the image as well as segments in subsequent images that have changed. These changes may be determined by determining one or more differences between the temporally stable depth map and segments included in one or more subsequent frames. Thereafter, the portions of the one or more subsequent frames that include segments including changes from their corresponding segments in the temporally stable depth map are processed and are combined with the segments from the temporally stable depth map to compute their associated disparities in one or more subsequent frames. The images may include a pair of stereo images acquired through a stereo camera system at a substantially similar time.

Abstract: Systems and methods of the present disclosure enable intelligent dynamic caching of data by accessing an activity history of historical electronic activity data entries associated with a user account, and utilizing a trained entity relevancy machine learning model to predict a degree of relevance of each entity associated with the historical electronic activity data entries in the activity history based at least in part on model parameters and activity attributes of each electronic activity data entry. A set of relevant entities are determined based at least in part on the degree of relevance of each entity. Pre-cached entities are identified based on pre-cached entity data records cached on the user device, and un-cached relevant entities from the set of relevant entities are identified based on the pre-cached entities. The cache on the user device is updated to cache the un-cached entity data records associated with the un-cached relevant entities.

Abstract: Provided are approaches for customizing transaction cards using a pressure sensitive modification area. In some approaches, a transaction card may include a body having a first main side opposite a second main side, wherein a chip is coupled to the first main side of the body, wherein a first outer layer extends along the first main side, and wherein a second outer layer extends along the second main side. The transaction card may further include an internal layer between the first and second outer layers, and a plurality of microspheres and a pigmented substance positioned between the internal layer and the first outer layer.