Abstract: Systems and methods are described for enhancing performance of a search engine by using pre-existing fine-tuned machine learning models that include domain-specific knowledge. A computing device receives search results from a search engine based on a query, in which the search results include chunk identifiers and a first weight score associated with each chunk identifier. The computing device further determines a subset of search results, which includes a first predetermined number of top-ranked chunk identifiers in the search results. The computing device generates, for each chunk identifier in the subset of search results, one or more second weight scores. Then the computing device generates, via a machine learning model, ensemble scores for the chunk identifiers. Afterwards, the computing device determines a second predetermined number of top-ranked chunk identifiers based on the ensemble scores.

Abstract: A method for reconciling data includes separating a first table into a plurality of first chunks, each first chunk including a plurality of first rows, and second table into a plurality of second chunks, each second chunk including a plurality of second rows. Each second chunk corresponds a first chunk. For each first chunk, the method includes: computing a first hash based on the first chunk and a second hash based on a second chunk corresponding to the first chunk; comparing the first hash with the second hash; in response to the first hash being different from the second hash, retrieving the plurality of first rows and the plurality of second rows and comparing the plurality of first rows to the plurality of second rows; in response to the first rows being different from the second rows, generating executable code configured to reconcile the first table with the second table.

Abstract: Methods and apparatuses for multimodal enhancement of interactions in conversation service applications include a server that establishes a chat-based communication session between a conversation service application and a client computing device. Query input is captured from a user during the session. The server converts an audio segment corresponding to a video into a video transcript using speech recognition, and determines a transcript portion for the video that is responsive to the query input using a dense passage retriever model. The server selects video frames from the video that correspond to the transcript portion and determines, for the selected frames, a predicted relevancy between the selected frame and the query input. The server generates a response to the query input comprising the selected frames based upon the relevancy. The server transmits the response to the client device for display.

Abstract: Methods and apparatuses are described for secure transmission and authentication of a user credential. A device comprising a memory, a processor, and a laser diode identifies a first user credential comprising a sequence of alphanumeric characters, converts the first user credential into a first plurality of analog signals, and activates the laser diode using the first plurality of analog signals to emit light detectable by a emitted light sensor of a second device. The second device comprising a memory, a processor, and an emitted light sensor generates a second plurality of analog signals corresponding to light emitted by the laser diode and detected by the emitted light sensor, converts the second plurality of analog signals into a second user credential, and authenticates the second user credential.

Abstract: The Address Verification, Seed Splitting and Firmware Extension for Secure Cryptocurrency Key Backup, Restore, and Transaction Signing Platform Apparatuses, Methods and Systems (“SFTSP”) transforms contract deployment request, transaction signing request, key backup request, key recovery request inputs via SFTSP components into contract deployment response, transaction signing response, key backup response, key recovery response outputs. A contract deployment request message datastructure is obtained. Owner key identification parameters are determined. An owner public key is determined using the owner key identification parameters. An owner address is generated using the owner public key. A salt value is generated. A contract address for the smart contract is calculated as a function of the deployment factory address, the salt value, the contract code, and the owner address. An owner private key is determined using the owner key identification parameters and used to sign the contract address.

Abstract: Systems and methods for extracting data from electronic documents using optical character recognition (OCR) and non-OCR based text extraction. A server computing device initiates non-OCR based text extraction for each page of an electronic document. The server calculates a document text coverage percentage corresponding to the non-OCR based text extraction for the whole document and, in response to determining that the document text coverage percentage is below a first threshold, initiates OCR for the document. The server calculates a page text coverage percentage corresponding to the non-OCR based text extraction for one or more pages of the electronic document and, in response to determining that the page text coverage percentage is below a second threshold, initiates OCR for the pages. The server combines first text extracted from the electronic document using non-OCR based text extraction and second text extracted from the electronic document using OCR.

Abstract: Methods and apparatuses for automated software container rehydration and deployment include a server that updates a container image for a software application by modifying layers of the container image and deploying the updated container image as a first container in a passive production environment. For a plurality of traffic volumes, the server: a) increases application traffic directed to the first container to a first traffic volume, b) monitors service level indicators associated with performance of the first container, c) increases the application traffic directed to the first container to a higher traffic volume upon determining that the service level indicators are within a performance threshold, and d) repeats steps b) and c) until the application traffic is increased to a highest traffic volume and the service level indicators are within the performance threshold. The server deploys the updated container image as a second container in an active production environment.

Abstract: Methods and apparatuses are described for automatic intelligent query suggestion for information retrieval applications. A server a) determines candidate intents associated with user input text received from a remote device, including applying a trained intent classification model to the user input text to predict candidate intents. The server b) calculates a likelihood value for each of the candidate intents. The server c) compiles a list of suggested queries based upon the candidate intents and associated likelihood values. The server d) identifies a subset of the list of suggested queries for display on the remote device. Upon detecting an update to the user input text at the remote device, the server repeats steps a) to d) using the updated user input text, or upon detecting a selection of one of the suggested queries at the remote device, the server retrieves content responsive to the selected query.

Abstract: Methods and apparatuses are described for automatic data-driven optimization of a target outcome using machine learning. A server generates a first feature dataset and applies a trained outcome prediction model to the first feature dataset as input to generate a second feature data set and a first predicted value for a target outcome. The server displays the first predicted value on a client device. The server receives input corresponding to one or more preferences or constraints from the client device and adjusts the trained outcome prediction model based upon the received input to incorporate the one or more preferences or constraints. The server applies the adjusted outcome prediction to the second feature dataset as input to generate a third feature data set a second predicted value for the target outcome. The server displays the second predicted value on the client device.

Abstract: Methods and apparatuses are described for automated cross-browser user interface testing. A computing device captures (i) a first image file corresponding to a first current user interface view of a web application on a first testing platform and (ii) a second image file corresponding to a second current user interface view of a web application on a second testing platform. The computing device prepares the image files, and compares the prepared image files using a structural similarity index measure. The computing device determines that the prepared first image file and the prepared second image file represent a common user interface view when the structural similarity index measure is within a predetermined range. The computing device highlights corresponding regions that visually diverge from each other in each of the prepared image files and transmits a notification message comprising the highlighted image files.

Abstract: Methods and systems are described for web object maintenance in a software application testing environment using reinforcement learning. A server detects failure of a test script caused by a web object of the software application. The server identifies the web object that caused the failure using code elements extracted from a webpage, including: determining, for each code element, actions to be performed against the code element, executing a deep neural network model to generate a reward value by applying each action to properties of the code element, selecting the code element having a maximum reward value, and classifying the code element upon comparing properties of the selected code element to properties of web objects in a repository. The server updates a web object in the repository to comprise the selected code element and the properties. The server resumes execution of the test script using the updated web object.

Abstract: Methods and apparatuses are described for provisioning and managing data orchestration platforms in a cloud computing environment. A server provisions in a first region a first data orchestration platform comprising (i) a first data transformation instance, (ii) first endpoints, and (iii) a first data integration instance. The server provisions in a second region a second data orchestration platform comprising (i) a second data transformation instance, (ii) second endpoints, and (iii) a second data integration instance. The server integrates the first data integration instance and the second data integration instance with an identity authentication service. The server monitors operational status of the first orchestration platform and the second orchestration platform using a monitoring service. The server refreshes virtual computing resources in each of the first orchestration platform and the second orchestration platform using a rehydration service.

Abstract: Methods and apparatuses are described for generating customized, context-specific visual artifacts using artificial intelligence (AI). A server computing device captures input data from one or more remote computing devices, the input data associated with one or more users. The server computing device creates one or more visual artifacts based upon the input data, each of the one or more visual artifacts associated with one or more of the users. The server computing device integrates the visual artifacts into a communication session associated with the remote computing devices.

Abstract: Methods and systems for secure geolocation-based data access control include a server that captures a data access request from a remote device and determines whether the requestor geolocation corresponds to a non-restricted zone or a restricted zone. The server determines whether the requestor identity has permission to receive a full view or a masked view of data. The server retrieves data responsive to the data query and generates a response to the data access request, the response including a full view or a masked view of the retrieved data. When the generated response comprises a masked view, the server determines that the user of the remote device has requested a full view of the responsive data, authenticates the remote device using the requestor identity, and updates the generated response to comprise the full view of the retrieved data. The server device transmits the generated response to the remote device.

Abstract: A computerized method is provided for extracting and visualizing data related to software development. The method includes receiving, by a computing device, a plurality of logs corresponding to respective ones of a plurality of code repositories associated with developing the software. Each code repository includes a Version Control System (VCS) configured to maintain the corresponding log and a plurality of code files associated with the development of the software over a time period. The method also includes extracting data from each log data, including dates of commit and commit messages over the time period associated with the code files. The method further includes analyzing the extracted data across the plurality of code repositories to determine information about one or more release events for the software and visualizing the one or more release events in a timeline view to graphically illustrate a development history of the software.

Abstract: Methods and apparatuses are described for disambiguating unrecognized abbreviations in search queries using machine learning. A server computing device detects unrecognized abbreviations in an initial search query received from a user of a client computing device. The server generates candidate words for each of the unrecognized abbreviations in the initial search query. For each of the abbreviations, the server executes a trained disambiguation machine learning model using the candidate words as input to generate, for each candidate word, a probability that the candidate word is intended by the user to be an expansion of the unrecognized abbreviation. The server selects one of the candidate words as a replacement word for the unrecognized abbreviation. The server generates a new search query by replacing each of the unrecognized abbreviations in the initial search query with the corresponding replacement word and retrieves search results by executing a search using the new search query.

Abstract: Automated intelligent correction of erroneous data for uninterrupted transaction processing includes a server that captures real-time transaction data from a transaction message stream. The server determines errors in the real-time transaction data for the transaction. The server identifies a trained artificial intelligence (AI) model from a plurality of trained AI models based upon the determined errors and executes the identified AI model using the real-time transaction data as input to correct the errors. The server determines a workflow state of the transaction using the real-time transaction data. The server generates a message for insertion in the transaction message stream, the message comprising the corrected real-time transaction data. The server inserts the generated message in the transaction message stream.

Abstract: Systems and methods for deriving classification rules from documents and a database using rule-based machine learning. The method includes extracting first variables from documents corresponding to an organization. The method further includes extracting second variables from a database corresponding to the organization. The method also includes filtering the extracted second variables based on at least one of null values, repeat variables, location variables, ID variables, or data variables. The method further includes deriving first classification rules based on the first variables using a rule-based machine learning algorithm. The method also includes calculating an accuracy of the derived first classification rules. The method also includes deriving second classification rules based on the first variables and the filtered second variables. The method further includes determining a suggested additional variable based on the derived second classification rules and the calculated accuracy.

Abstract: Methods and apparatuses are described for generating a digital image using an unencrypted key. A server computing device generates an initial digital image comprising a plurality of image areas. The server computing device converts an unencrypted key associated with a remote computing device into an encrypted key using an encryption algorithm. The server computing device assigns a color value to each character of the encrypted key. The server computing device maps each character of the encrypted key to one of the plurality of image areas of the initial digital image. The server computing device generates an updated digital image by applying, to each of the plurality of image areas, a color that corresponds to the color value assigned to the character of the encrypted key that is mapped to the image area. The server computing device stores the updated digital image in a database record.

Abstract: A computer-implemented method is provided for quantitative performance evaluation of a call agent. The method comprises converting an audio recording of a call between the call agent and a customer to a text-based transcript and identifying at least one topic for categorizing the transcript. The method also includes retrieving a set of criteria associated with the topic. Each criterion correlates to a set of predefined questions for interrogating the transcript to evaluate the performance of the call agent with respect to the corresponding criterion. Each question captures a sub-criterion under the corresponding criterion. The method further includes inputting the predefined questions and the transcript into a trained large language model to obtain scores for respective ones of the predefined questions. Each score measures a degree of satisfaction of the performance of the call agent during the call with respect to the sub-criterion captured by the corresponding predefined question.