IP Library Granted Patent US 12670048
Granted Patent B2
US 12670048 · App. 18/805,012 · Granted Jun 30, 2026

Systems and methods for recovery strategy creation, evaluation, and recommendation

Inventors: Eric R Alspaugh (Lancaster, OH); Archana Ramesh (Bengaluru, IN)
Assignee: JPMORGAN CHASE BANK, N.A.
G06F11/008G06F11/0751G06F11/079G06F11/0793
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12670048
App. No.
18/805,012
Granted
Jun 30, 2026
Kind
B2
Abstract

In some aspects, the techniques described herein relate to a method comprising determining an event associated with a threshold probability of causing inaccessibility to an asset; retrieving, by a model executed by a processor and from a database, event data; determining, from a database of a plurality of assets, a set of assets affected by the event based on the event data; generating cause synthetic data related to the event for the set of assets to determine a category for the event based on a location of the event; generating effect synthetic data related to the event for the set of assets; generating a recovery strategy including a response detail based on the cause synthetic data and the effect synthetic data for each asset; generating a risk profile including the response detail for each asset and a probability of the event; and calculating a degree of risk.

Claims (19)

1 . A method comprising:

determining an event associated with a threshold probability of causing inaccessibility to an asset;

retrieving, by a model executed by a processor and from a database, event data, wherein the event data includes one or more of an event category, a nature, a location, a scale, a duration, a season, and a time of day;

determining, from a database of a plurality of assets, a set of assets affected by the event based on the event data;

generating, by a generative adversarial network and a probabilistic time series model, cause synthetic data related to the event for the set of assets to determine a category for the event based on a location of the event, wherein the cause synthetic data includes inputs of past event data with the same impact, time, or nature, wherein the generative adversarial network determines whether a generated event data of the cause synthetic data meets a threshold metric based on comparison to actual event data;

generating effect synthetic data related to the event for the set of assets to determine a first order impact and a second order impact;

generating a recovery strategy including a response detail based on the cause synthetic data and the effect synthetic data for each asset;

generating a risk profile associated with the generated recovery strategy including the response detail for each asset and a probability of the event; and

calculating a degree of risk for the response detail.

2 . The method of claim 1 , wherein the effect synthetic data is generated by the generative adversarial network including inputs of a mix of the set of assets and a probabilistic time series model to determine whether the generated event data meets a threshold metric based on comparison to actual complex asset disruption data.

3 . The method of claim 1 , wherein the recovery strategy is generated by the generative adversarial network including inputs of the cause synthetic data and the effect synthetic data to determine whether the recovery detail meets a threshold metric of recovery of an asset of the set of assets, wherein the asset of the set of assets is impacted by the event.

4 . The method of claim 1 , further comprising generating, with a retrieval augment generation model, chunks based on the recovery strategy.

5 . The method of claim 4 , further comprising combining the chunks with a query to construct a prompt.

6 . The method of claim 1 , wherein generating the cause synthetic data or the effect synthetic data comprises generating a chaotic event.

7 . The method of claim 5 , further comprising using retrieved relevant documents as context with the prompt and the query and generating a response with a language learning model.

8 . The method of claim 1 , further comprising querying through a prompt systematically to recommend one or more ways for determining the response detail for the recovery strategy using a combination of parameters.

9 . The method of claim 1 , further comprising querying through a prompt systematically to recommend one or more ways for determining the response detail for the recovery strategy using a confidence score.

10 . The method of claim 1 , further comprising enacting a response by allocating an asset to the response detail.

11 . The method of claim 10 , further comprising, based on the allocation, contacting an asset with a task related to the response detail.