Abstract: Systems, computer program products, and methods are described herein for machine learning-based metadata collection and parameterized graph modeling from communication channels. The present disclosure comprises a communication interaction subsystem (CIS) configured to receive requests from a user input device to query network traffic data associated with a plurality of devices. The request comprises a factor set and a correlation criteria. The CIS analyzes the network traffic data based on at least the request and determines a subset of the plurality of devices based on at least the request. The system also comprises a parameterized graph modeling subsystem (PGMS) operatively coupled to the CIS, which is configured to generate a data traffic topography map associated with the subset of the plurality of devices and transmit control signals configured to cause the user input device to display the data traffic topography map.

Abstract: Arrangements for dynamic identity confidence modeling are provided. In some aspects, first identity information associated with the user may be received. An identity confidence model associated with the user indicating a level of confidence that the user is authentic may be generated. User activity data associated with transactions and interactions of the user may be received. The user activity data may include temporal information associated with the user transacting and interacting with an entity at one or more touchpoints. Second identity information associated with the user may be extracted and compared to the first identity information using machine learning. One or more anomalies may be identified and authentication information associated with the identified one or more anomalies may be requested. The identity confidence model associated with the user may be automatically and continuously updated based at least in part on the comparison.