# US 2026/0094017 A1 · Dynamically generating knowledge assessment items

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0094017 A1
- Document ID: US20260094017A1
- Status: Published application
- Kind: A1
- Application number: 19/315,716
- Filed: 2025-09-01
- Published: 2026-04-02
- Earliest claimed: 2024-09-04
- Assignees: Examroom.Ai Corp.
- Applicants: Examroom.Ai Corp.
- Inventors: Vinod Kumar JAYAKEERTHI; Parshant KUMAR; Ravichandra PANCHAKSHARAM; Deepak Madhukar KOLEKAR
- Web page: https://patentgrep.com/patent/US20260094017A1

## CPC

- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Embodiments of the present disclosure relate to a method, a system and a computer program product for generating and/or creating knowledge assessment items for an item bank using artificial intelligence and machine learning.

## Claims

### Claim 1

A method for dynamically generating knowledge assessment items, the method comprising:
an artificial intelligence/machine learning (AI/ML) module, the AI/ML module configured for
receiving as input from a source at least one or more knowledge assessment item type and a subject area, wherein knowledge assessment items are generated from the input provided to the AI/ML module, wherein the input is at least in one a structured format, an unstructured format and a combination thereof;
based on the input provided to the AI/ML module, the AI/ML module configured for:
fetching relevant content from at least one or more content sources relevant to the subject area;
dynamically creating at least one or more knowledge assessment items in the one or more knowledge assessment item type, wherein the at least one or more knowledge assessment items output comprises at least one of a dichotomous type and a polychotomous type; and
providing as output from the AI/ML module the at least one or more knowledge assessment items created to the source.

### Claim 2 (depends on 1)

The method of claim 1, wherein content is fetched from at least one of a local source or a repository or an internet site or an intranet site or a book or a journal or a multi-media source or an audio source or a video source or an audio-visual source or a combination thereof.

### Claim 3 (depends on 1)

The method of claim 1, wherein the source comprises at least one of an examination administrator, a client, a third party acting on behalf of the client, an authorized person acting on behalf of the client and a subject matter expert.

### Claim 4 (depends on 1)

The method of claim 1, wherein the knowledge assessment items generated by the AI/ML module are authenticated and vetted by the source prior to updating the knowledge assessment items to an item bank.

### Claim 5 (depends on 4)

The method of claim 4, wherein at least one of the AI/ML module or the source assigns a difficulty score for each of the knowledge assessment items generated, wherein at least one of the AI/ML module automatically assigns a difficult score to the knowledge assessment item, the source uses human knowledge with expertise to assign a difficulty score to the knowledge assessment item, the source uses human knowledge with expertise and guideline provided by the AI/ML module to assign a difficulty score to the knowledge assessment item and the difficulty score assigned by the AI/ML module can be overridden by the source assigning a new difficulty score.

### Claim 6 (depends on 5)

The method of claim 5, wherein the difficulty score assigned to a knowledge assessment item varies for different item type, wherein the difficulty score for the knowledge assessment item in a first item type varies from the difficulty score for the knowledge assessment item in a second item type, wherein the first item type is different from the second item type.

### Claim 7 (depends on 1)

The method of claim 1, wherein fetching content comprises at least one of:
automatically fetching content in the relevant subject area, and requesting the source to manually provide the content in the relevant subject area; and
preprocessing the content by the AI/ML module to validate the content.

### Claim 8 (depends on 1)

The method of claim 1, wherein the knowledge assessment item comprises at least a stem, at least a set of keys, at least a set of distractors, and at least an input requirement to be provided by a candidate for a stem.

### Claim 9 (depends on 1)

The method of claim 1, further comprising eliminating duplicates items generated in at least one of a same item type category, the dichotomous item types and the polychotomous item types.

### Claim 10 (depends on 1)

The method of claim 1, wherein the AI/ML module comprises performing at least one of a Retrieval-Augmented Generation (RAG) framework, a Knowledge Augmented Generation (KAG) framework, a Model context protocol (MCP) framework, a generative AI framework, a transformer model framework. a deep learning model, an elastic weight consolidation (EWC) model, an instructor-based training model, progressive neural network model, a learning without forgetting model, a memory replay system model, a vector-based model, graph-based model and McCulloch-Pitts Neuron framework, agentic framework, Knowledge Distillation framework, Transfer Learning framework, and Reinforcement learning framework for generating items.

### Claim 11

A system comprising at least a processor and a memory, the system further comprises an artificial intelligence/machine learning (AI/ML) module, the AI/ML module configured for
receiving as input from a source at least one or more knowledge assessment item type and a subject area, wherein knowledge assessment items are generated from the input provided to the AI/ML module, wherein the input is at least in one a structured format, an unstructured format and a combination thereof;
based on the input provided to the AI/ML module, the AI/ML module configured for:
fetching relevant content from at least one or more content sources relevant to the subject area;
dynamically creating at least one or more knowledge assessment items in the one or more knowledge assessment item type, wherein the at least one or more knowledge assessment items output comprises at least one of a dichotomous type and a polychotomous type; and
providing as output from the AI/ML module the at least one or more knowledge assessment items created to the source.

### Claim 12 (depends on 11)

The system of claim 11, wherein content is fetched from at least one of a local source or a repository or an internet site or an intranet site or a book or a journal or a multi-media source or an audio source or a video source or an audio-visual source or a combination thereof.

### Claim 13 (depends on 11)

The system of claim 11, wherein the source comprises at least one of an examination administrator, a client, a third party acting on behalf of the client, an authorized person acting on behalf of the client and a subject matter expert.

### Claim 14 (depends on 11)

The system of claim 11, wherein the knowledge assessment items generated by the AI/ML module are authenticated and vetted by the source prior to updating the knowledge assessment items to an item bank.

### Claim 15 (depends on 14)

The system of claim 14, wherein at least one of the AI/ML module or the source assigns a difficulty score for each of the knowledge assessment items generated, wherein at least one of the AI/ML module automatically assigns a difficult score to the knowledge assessment item, the source uses human knowledge with expertise to assign a difficulty score to the knowledge assessment item, the source uses human knowledge with expertise and guideline provided by the AI/ML module to assign a difficulty score to the knowledge assessment item and the difficulty score assigned by the AI/ML module can be overridden by the source assigning a new difficulty score.

### Claim 16 (depends on 15)

The system of claim 15, wherein the difficulty score assigned to a knowledge assessment item varies for different item type, wherein the difficulty score for the knowledge assessment item in a first item type varies from the difficulty score for the knowledge assessment item in a second item type, wherein the first item type is different from the second item type.

### Claim 17 (depends on 11)

The system of claim 11, wherein fetching content comprises at least one of:
automatically fetching content in the relevant subject area, and requesting the source to manually provide the content in the relevant subject area;
preprocessing the content by the AI/ML module to validate the content.

### Claim 18 (depends on 11)

The system of claim 11, wherein the knowledge assessment item comprises at least a stem, at least a set of keys, at least a set of distractors, and at least an input requirement to be provided by a candidate for a stem.

### Claim 19 (depends on 1)

The system of claim 1, further comprising eliminating duplicates items generated in at least one of a same item type category, the dichotomous item types and the polychotomous item types.

### Claim 20 (depends on 11)

The method of claim 11, wherein the AI/ML module comprises performing at least one of a Retrieval-Augmented Generation (RAG) framework, a Knowledge Augmented Generation (KAG) framework, a Model context protocol (MCP) framework, a generative AI framework, a transformer model framework. a deep learning model, an elastic weight consolidation (EWC) model, an instructor-based training model, progressive neural network model, a learning without forgetting model, a memory replay system model, a vector-based model, graph-based model and McCulloch-Pitts Neuron framework, agentic framework, Knowledge Distillation framework, Transfer Learning framework, and Reinforcement learning framework for generating items.
