# US 2025/0157670 A1 · Ensemble machine learning systems and methods

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

- Number: US 2025/0157670 A1
- Document ID: US20250157670A1
- Status: Published application
- Kind: A1
- Application number: 19/026,469
- Filed: 2025-01-17
- Published: 2025-05-15
- Earliest claimed: 2018-04-06
- Assignees: Curai, Inc.
- Applicants: Curai, Inc.
- Inventors: Anitha Kannan; Murali Ravuri; Vitor Rodrigues; Vignesh Venkataraman; Geoffrey Tso; Neal Khosla; Neil Hunt; Xavier Amatriain; Manish Chablani
- Web page: https://patentgrep.com/patent/US20250157670A1

## CPC

- G16H 10/20: for electronic clinical trials or questionnaires
- G16H 50/30: for calculating health indices; for individual health risk assessment
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/09: Supervised learning
- G06N 3/096: Transfer learning
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 5/027: {Frames}
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- G06N 20/20: Ensemble learning
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G16H 20/00: ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H 40/67: for remote operation
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G16H 70/60: relating to pathologies
- G16H 80/00: ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/08: Learning methods
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Techniques for responding to a healthcare inquiry from a user are disclosed. In one particular embodiment, the techniques may be realized as a method for responding to a healthcare inquiry from a user, according to a set of instructions stored on a memory of a computing device and executed by a processor of the computing device, the method comprising the steps of: classifying an intent of the user based on the healthcare inquiry; instantiating a conversational engine based on the intent; eliciting, by the conversational engine, information from the user; and presenting one or more medical recommendations to the user based at least in part on the information.

## Claims

### Claim 1

A computer-based system comprising an ensemble of machine learning models configured to provide healthcare to a user, the computer-based system comprising:
(a) a conversational machine learning model comprising a language model that is trained on a corpus of text and configured to carry out, or assist in carrying out, a conversation with the user comprising a healthcare interaction;
(b) a diagnostic machine learning model that is configured to generate one or more of a differential diagnosis for the user and a diagnosis for the user based, at least in part, on the conversation; and
(c) a task machine learning model that is configured to carry out a healthcare task for the user based on one or more of the conversation, the differential diagnosis, and the diagnosis;
wherein the conversational machine learning model, the diagnostic machine learning model, and the task machine learning model form the ensemble of machine learning models.

### Claim 2 (depends on 1)

The computer-based system of claim 1, wherein the computer-based system is configured to receive multimodal input from the user.

### Claim 3 (depends on 2)

The computer-based system of claim 2, wherein the multimodal input comprises one or more of written input, spoken input, non-speech audio input, static image input, and video input.

### Claim 4 (depends on 1)

The computer-based system of claim 1, wherein the conversation with the user includes at least one input from the user that is unstructured.

### Claim 5 (depends on 4)

The computer-based system of claim 4, wherein natural language processing is applied to the at least one input from the user that is unstructured.

### Claim 6 (depends on 1)

The computer-based system of claim 1, wherein the language model comprises a deep machine learning model.

### Claim 7 (depends on 1)

The computer-based system of claim 1, wherein the conversational machine learning model is configured to minimize a number of questions needed to ask the user in the conversation to generate the one or more of the differential diagnosis and the diagnosis.

### Claim 8 (depends on 1)

The computer-based system of claim 1, wherein the diagnostic machine learning model is at least partially trained using computer-generated medical data that is produced by a generative model.

### Claim 9 (depends on 1)

The computer-based system of claim 1, wherein the healthcare task comprises connecting the user with a healthcare provider.

### Claim 10 (depends on 1)

The computer-based system of claim 1, wherein the healthcare task comprises one or more of issuing a prescription, renewing a prescription, and providing a referral.

### Claim 11 (depends on 1)

The computer-based system of claim 1, wherein the healthcare task comprises arranging testing for the user.

### Claim 12 (depends on 1)

The computer-based system of claim 1, wherein the healthcare task comprises providing medical advice related to one or more of the conversation, the differential diagnosis, and the diagnosis.

### Claim 13 (depends on 12)

The computer-based system of claim 12, wherein the medical advice that is provided is regulated by a regulatory authority.

### Claim 14

A computer-based method for utilizing an ensemble of machine learning algorithms to provide healthcare to a user, the method comprising:
carrying out, with a conversational machine learning model, comprising a language model that is trained on a corpus of text, a computer-generated conversation with the user related to a health of the user;
applying a reasoning machine learning model to at least a portion of the conversation to determine a medical advice to provide to the user; and
providing the medical advice to the user;
wherein the conversational machine learning model and the reasoning machine learning model form the ensemble of machine learning models.

### Claim 15 (depends on 14)

The computer-based method of claim 14, wherein the language model comprises a deep machine learning model.

### Claim 16 (depends on 14)

The computer-based method of claim 14, wherein the medical advice comprises medical advice that is regulated by a regulatory authority.

### Claim 17 (depends on 14)

The computer-based method of claim 14, comprising generating, by the conversational machine learning model, text that elicits an input from the user.

### Claim 18 (depends on 14)

The computer-based method of claim 14, comprising generating a classification of an input from the user.

### Claim 19 (depends on 18)

The computer-based method of claim 18, comprising receiving, by the reasoning machine learning model of the ensemble of machine learning models, at least one of the input from the user that is elicited by the conversational machine learning model and the classification of the input of the user that is generated.

### Claim 20 (depends on 14)

The computer-based method of claim 14, wherein the ensemble of machine learning algorithms is configured to estimate or determine an objective of the user based on one or more inputs of the user within the conversation.

### Claim 21 (depends on 20)

The computer-based method of claim 20, comprising finding, using the conversational machine learning model, a minimum number of questions to ask the user during the conversation to achieve the objective of the user.

### Claim 22 (depends on 14)

The computer-based method of claim 14, comprising eliciting, using the conversational machine learning model, input types that will lead to a final disposition with respect to the user using a fewest amount words or sentences.

### Claim 23 (depends on 14)

The computer-based method of claim 14, comprising creating, using the conversational machine learning model, a profile of the user based on the conversation.

### Claim 24 (depends on 23)

The computer-based method of claim 23, comprising generating an understanding of an input by the user based at least in part on the profile of the user.

### Claim 25 (depends on 14)

The computer-based method of claim 14, comprising generating a written document based on one or more of the conversation and the medical advice.

### Claim 26 (depends on 25)

The computer-based method of claim 25, wherein the written document that is generated comprises a computer-generated physician note or a portion of a physician note.

### Claim 27 (depends on 14)

The computer-based method of claim 14, wherein the conversational machine learning model is trained one or more recorded conversations between a patient and a physician.

### Claim 28 (depends on 14)

The computer-based method of claim 14, comprising carrying out a subsequent conversation with the user, after the medical advice is provided.

### Claim 29 (depends on 28)

The computer-based method of claim 28, comprising determining an outcome related to one or more of the user and the medical advice based on the subsequent conversation.

### Claim 30 (depends on 29)

The computer-based method of claim 29, wherein the outcome is used to further train the conversational machine learning model.
