Abstract: A system includes a processor and a memory including instructions that are executable for causing the processor to perform operations. The operations include accessing a new component value and a set of inner heaps. Each inner heap of the set of inner heaps includes a sorted set of components originating from a plurality of component sources. The operations further include assigning the new component value to a first inner heap of the set of inner heaps. In response to assigning the new component value to the first inner heap, the operations include sorting the first inner heap to identify a maximum or minimum value of the first inner heap. In response to determining that the new component value changes the maximum or minimum value of the first inner heap, the operations include updating an outer heap to include a reference to the new component value of the first inner heap.

Abstract: Systems and methods for authenticating objects associated with online activities are provided. For a target object involved in online activities, an authentication system accesses data associated with the target object and generates a featurized dataframe for the target object by processing the data associated with the target object. The authentication system further applies a machine learning model to the featurized dataframe to determine an authenticity indicator for the target object and assigns the target object to an authentication stream according to the authenticity indicator determined. Based on the results returned from the assigned authentication stream, the online activities involving the target object may be blocked or permitted.