Artificial intelligence system augmented with an expert system scaled with generative artificial intelligence

1 . A computer implemented method for generating an application, the method comprising:

receiving a category that specifies a class, or group, of items that comprise things or people that share a commonality or a common attribute;

providing an option to generate criteria of the category by,

providing the category to a generative artificial intelligence service that provides in response to the category a first listing of criteria of the category; and

retrieving a second listing of criteria of the category from storage;

receiving one of the first listing of criteria of the category and the second listing of criteria of the category as a selected listing of criteria;

selecting certain criterion of the selected listing of criteria to generate a subset listing of criteria;

scoring the subset listing of criteria to generate a scored listing comprising a score associated with each criterion in the subset listing of criteria;

analyzing the scored listing by identifying data types and data patterns and distribution in each criterion of the scored listing and storing the results as a structured data set in a structured data format; and

generating the application by retrieving the structured data set and generating a user interface by which a user of the application provides a user selected category and in response receives an ordered set of items within the user selected category ordered by criteria of the category.

2 . The computer implemented method of claim 1 further comprising scoring each item in the ordered set of items in accordance with criteria of the category with a machine-learning engine.

3 . The computer implemented method of claim 1 further comprising scoring each item in the ordered set of items in accordance with criteria of the category with a rule-based scoring engine.

4 . The computer implemented method of claim 1 wherein the user interface of the application receives for the criteria of the user selected category, tradeoffs from the user expressed in a form of relative weights.

5 . The computer implemented method of claim 2 wherein the operation of scoring each item in the ordered set of items in accordance with criteria of the category with a machine-learning engine comprises forming a request that complies with an application programming interface of the machine-learning engine and receiving from the machine-learning engine scores for each item in the ordered set of items.

6 . The computer implemented method of claim 1 further comprising scoring each item in the ordered set of items in accordance with criteria of the category by retrieving the scoring of each item in the ordered set of items from a database.

7 . The computer implemented method of claim 1 wherein the operation of scoring the subset listing of criteria to generate a scored listing comprising a score associated with each criterion in the subset listing of criteria comprises:

providing the subset listing of criteria to an artificial intelligence/machine learning (AI/ML) engine with each criterion of the subset listing of criteria corresponding to a label for the AI/ML engine, wherein the AI/ML engine is trained to provide the score associated with each criterion of the subset listing of criteria.

8 . The computer implemented method of claim 7 wherein the AI/ML engine comprises a plurality of sub-AI/ML engines where each sub-AI/ML engine is trained for the category corresponding to the subset listing of criteria.

9 . The computer implemented method of claim 1 wherein the operation of scoring the subset listing of criteria to generate a scored listing comprising a score associated with each criterion in the subset listing of criteria comprises, providing the subset listing of criteria to a function module to select one of a plurality of functions.

10 . The computer implemented method of claim 1 wherein the operation of analyzing the scored listing by identifying data types and data patterns and distribution in each criterion of the scored listing and storing the results as a structured data set in a structured data format comprises:

identifying for each criterion a corresponding criterion behavior by identifying a data type, data patterns and data distribution; and

generating a set of utility curves for each criterion that correspond to the criterion behavior.

11 . A computer-implemented method that operates to generate tradeoff-based decisions by determining weighted preference information including a plurality of decision criteria obtained from a generative AI engine from a corresponding plurality of weights representing relative importance of each criterion, wherein determining weighted preference information comprises:

obtaining a category;

obtaining for the category a set of criteria by providing the category to a generative artificial intelligence (AI) engine, and receiving the set of criteria as results from the generative AI engine;

identifying a function corresponding to each criterion of the set of criteria and reflecting the criterion behavior;

using the function to quantify the unweighted contribution of each option for each corresponding criterion;

providing a set of weights either explicitly or implicitly by inference; and

yielding a tradeoff-based decision by ranking the results based upon the weighted preference information.

12 . A server computer system that interacts with an application executing on a client computer system, the server computer system comprising:

data storage having stored therein a database that stores information pertaining to items, where each item corresponds to a category, wherein the database includes one or more categories of items;

a processor that executes instructions that cause the processor to:

provide to the application a ranking interface,

that receives from the application a first category and application settings in a form of tradeoffs; and

applies the tradeoffs to items in the database that correspond to the first category to provide to the application a ranked listing of the items from the database, where each item provided is accompanied with a corresponding score that indicates correspondence of the item to the received tradeoffs;

provide to the application an inference interface that,

receives from the application a second category and an ordered list of choices within the category that indicate a user preference;

infers tradeoff values to match an ordering of choices; and

provides to the application a listing of criteria together with a weight for each criteria, wherein the weights correspond to the user preference.

13 . The server computer system of claim 12 wherein the processor applies the tradeoffs to items in the database that correspond to the first category to provide to the application a ranked listing of the items from the database, where each item provided is accompanied with a corresponding score that indicates correspondence of the item to the received tradeoffs by:

generating, as a function of the first category and the tradeoffs, a query to a generative artificial intelligence engine for an initial set of criteria that characterize the first category, wherein the initial set of criteria comprises a plurality of criterion characterized by an ordering;

receiving the initial set of criteria and identifying a utility curve that substantially fits the initial set of criteria; and

applying the utility curve to items in the database to items in the database that correspond to the first category to provide to the application a ranked listing of the items from the database.

14 . The server computer system of claim 12 wherein the processor infers tradeoff values to match the ordering of choices by reverse propagation using a ranking of a subset of items in the second category.

15 . The server computer system of claim 12 wherein the processor provides to the application a listing of criteria together with a weight for each criterion, wherein the weights correspond to the user preference, by:

retrieving from a database a set of items that have associated with them an ordering and using the tradeoff values to match the ordering of items and providing to the application a ranked listing of the items from the database, where each provided item is provided with a corresponding score that indicates its correspondence to the inferred tradeoffs.