Abstract: An apparatus and method for providing an immutable audit trail for machine learning applications is described herein. The audit trail is preserved by recording the machine learning models and data in a data structure in immutable storage such as a WORM device, a cloud storage facility, or in a blockchain. The immutable audit trail is important for providing bank auditors with the reasons for lending or account opening reasons, for example. A graphical user interface is described to allow the archive of machine learning models to be viewed.

Abstract: A virtual fraud detection system and method is described for the real time processing of banking transactions seen on a banking rail. The transaction is processed through natural language processing to determine who the parties are, and natural language processing is performed on the web site and the social media pages of employees to ascertain if the originator and the beneficiary of the transaction make sense. In addition, the age of the DNS records of the parties is checked to see if the parties are established organizations.

Abstract: Disclosed is a system for scoring customers of a financial institution based on financial data. The system includes a central database that stores a plurality of modules, a central server that processes the plurality of modules and a display unit that displays the processed plurality of modules. The plurality of modules includes a criteria configuration module, a data module, and a computation module. The criteria configuration module includes a metric module to receive the input parameters required to evaluate the score, and a measurement module for defining transformation criteria to be applied on the data corresponding to the input parameters. The computation module includes a metric evaluation module to compute and applies the transformation criteria to the values of the input parameters, and a scoring module coupled to the metric evaluation module to automatically compute and display the score of the customers based on the values retrieved from the metric evaluation module.

Abstract: An improved apparatus and method for detecting fraud is described using a stacked auto encoder with embedding to encode and decode a transaction to determine fraud. The technique includes model tuning software and transaction review software. The model tuning software views the transaction and tunes an artificial neural network model to minimize reconstruction loss. The transaction review software processes the transaction through the artificial neural network model, converting the transaction into a feature vector, encoding the feature vector into a compressed vector, decoding the compressed vector into a reconstructed vector, subtracting the reconstructed vector from the feature vector, and determining a fraud indication and reasoning based on a difference from the reconstructed vector from the feature vector.

Abstract: A method for creating machine learning model performance alerts showing the drifting of functions is described herein. The method starts by creating the initial machine learning model using a training data set. This initial machine learning model is then used in production, and the model is updated to account for the production data. To assure the quality of the updated machine learning model, test data results from the initial machine learning model is compared to the results from the updated machine learning model. Each feature is checked to see if the difference is within a p-value and whether the confidence intervals overlap. If not, an alert is generated to take action on the model.

Abstract: A method and apparatus for detecting suspicious activities surrounding the management of controlled substances in a medical facility is described herein, where the activities that may indicate the diversion of controlled drugs are flagged for further review. The activities are detected by reviewing application layer network packets, related to controlled substances, on the medical facility network, identifying the physical location of the origin of the packet, and processing these packets with a rules engine and machine learning generated rules to make a determination if the circumstances surrounding the packets indicate the diversion of controlled substances.

Abstract: A unique user interface for improving machine learning algorithms is described herein. The user interface comprises an icon with multiple visual indicators displaying the machine learning confidence score. When a mouse hovers over the icon, a set of icons are displayed to accept the teaching user's input. In addition, the words that drove the machine learning confidence score are highlighted with formatting so that the teaching user can understand what drove the machine learning confidence score.

Abstract: A two-step algorithm for conducting near real time fuzzy searches of a target on one or more large data sets is described, where the address and the geolocation are included in the match criteria. This algorithm includes the simplification of the data by removing grammatical constructs to bring the target search term (and the stored database) to their base elements and then perform a Levenshtein comparison to create a subset of the data set that may be a match. Location is determined by a geohash comparison of the latitude and longitude and a Levenshtein comparison of the address text. Then performing a scoring algorithm while comparing the target to the subset of the data set to identify any matches.

Abstract: A two-step algorithm for conducting near real-time fuzzy searches of a target on one or more large data sets is described. This algorithm includes the simplification of the data by removing grammatical constructs to bring the target search term (and the stored database) to their base elements and then perform a Levenstein comparison to create a subset of the data set that may be a match. Then performing a scoring algorithm while comparing the target to the subset of the data set to identify any matches.

Abstract: Disclosed is a system for authenticating checks using a mixed reality environment. The system includes a financial instrumentation central database for storing financial and identification data related to the checks, and a reality glasses coupled to the financial instrumentation central database. The reality glass includes a memory unit for storing a plurality of modules, a camera to capture the image of the check, a reality display; and a wearable processor coupled to the memory unit for processing the stored plurality of modules. The wearable processor is coupled to the financial instrumentation central database, to the camera to process captured images of the check.

Abstract: Disclosed is a system for facilitating financial planning of an objective. The system includes a storage unit for storing computer program instructions, a display unit for displaying processed computer program instructions, a processing unit is coupled to the storage unit and the display unit for processing the computer program instructions. The computer program instructions includes a data input computer program instruction, a data category display computer program instruction, a threshold computer program instruction, a slide computer program instruction and a heat map computer program instruction. The heat map computer program instruction with a color computer program instruction to display the degree of difference with the intensity of the color. A red color to display the degree of difference for a negative value and a green color to display the degree of difference for a positive value.

Abstract: A computer-implemented method for providing tenant aware, variable length, deduplication of data stored on a non-transitory computer readable storage medium is described here. The method is performed at least in part by circuitry and the data comprises a plurality of data items. Each of the plurality of data items is associated with a particular tenant of a group of tenants that store data on the storage medium. In addition, a data deduplication data storage medium made up of one or more tenant storage areas and a deduplicated block database stored on the data deduplication data storage medium is described. The tenant storage areas comprise a plurality of data items, where each data item comprises a plurality of item blocks and each item block comprises a pointer to a storage location in a deduplicated block database.

Abstract: A two-step algorithm for conducting near real time fuzzy searches of a target on one or more large data sets is described, where the address and the geolocation are included in the match criteria. This algorithm includes the simplification of the data by removing grammatical constructs to bring the target search term (and the stored database) to their base elements and then perform a Levenshtein comparison to create a subset of the data set that may be a match. Location is determined by a geohash comparison of the latitude and longitude and a Levenshtein comparison of the address text. Then performing a scoring algorithm while comparing the target to the subset of the data set to identify any matches.

Abstract: An improved method and apparatus for determining if a financial transaction is fraudulent is described. The apparatus in one embodiment collects transactions off of a rail using promiscuous listening techniques. The method uses linear programming algorithms to tune the rules used for making the determination. The tuning first simulates using historical data and then creates a matrix of the rules that are processed through the linear programming algorithm to solve for the variables in the rules. With the updated rules, a second simulation is performed to view the improvement in the performance. The updated rules are then used to evaluate the transactions for fraud.

Abstract: A two-step algorithm for conducting near real-time fuzzy searches of a target on one or more large data sets is described. This algorithm includes the simplification of the data by removing grammatical constructs to bring the target search term (and the stored database) to their base elements and then perform a Levenstein comparison to create a subset of the data set that may be a match. Then performing a scoring algorithm while comparing the target to the subset of the data set to identify any matches.

Abstract: A computer-implemented method for providing tenant aware, variable length, deduplication of data stored on a non-transitory computer readable storage medium. The method is performed at least in part by circuitry and the data comprises a plurality of data items. Each of the plurality of data items is associated with a particular tenant of a group of tenants that store data on the storage medium.