US 11,989,217 B1
Systems and methods for real-time data processing of unstructured data
Miriam Silver, Tel Aviv (IL); Nimrod Barak, New York, NY (US); Prag Sharma, New York, NY (US); and Avi Gefen, New York, NY (US)
Assigned to Citibank, N.A., New York, NY (US)
Filed by Citibank, N.A., New York, NY (US)
Filed on Dec. 6, 2023, as Appl. No. 18/531,660.
Int. Cl. G06F 16/33 (2019.01); G06F 16/31 (2019.01); G06F 40/40 (2020.01)
CPC G06F 16/3347 (2019.01) [G06F 16/31 (2019.01); G06F 40/40 (2020.01)] 20 Claims
OG exemplary drawing
 
1. A system for providing real-time, user-specific notifications of unstructured data by processing the unstructured data without interstitial standardization, the system comprising:
receiving a first native unstructured dataset;
determining a first native unstructured dataset component in the first native unstructured dataset;
generating a first vector representation of the first native unstructured dataset component;
generating a vector database comprising the first vector representation;
generating a first pointer to the first native unstructured dataset component in the first native unstructured dataset;
processing the vector database to determine, using a first artificial intelligence model, a first dependency of the first native unstructured dataset component, wherein the first dependency comprises a third native unstructured dataset component in a second native unstructured dataset, and wherein the first artificial intelligence model is trained to determine dependencies between native unstructured datasets based on historic vector representations of historic unstructured datasets;
processing the vector database to determine, using a second artificial intelligence model, first content for the first native unstructured dataset component, wherein the second artificial intelligence model uses natural language processing to determine the first content;
determining an urgency of a first notification based on a first temporal identifier in the first content;
determining a user of the first notification based on a first user identifier in the first content; and
generating for display, on a user interface, the first notification based on the user and the urgency, wherein the first notification is populated based on retrieving the first native unstructured dataset component using the first pointer and retrieving the third native unstructured dataset component based on the first dependency.