System and method for providing model-based treatment recommendation via individual-specific machine learning models

Patent metadata

Field	Value
Publication number	US-10896763-B2
Application number	US-201916242350-A
Country	US
Kind code	B2
Filing date	Jan 8, 2019
Priority date	Jan 12, 2018
Publication date	Jan 19, 2021
Grant date	Jan 19, 2021

Abstract

The present disclosure pertains to a system for providing model-based treatment recommendation via individual-specific machine learning models. In some embodiments, the system (i) obtains an audio recording of an individual, (ii) determines, from the audio recording, one or more utterance-related features of the individual; (iii) performs one or more queries based on the one or more utterance-related features to obtain health information (e.g., utterance-related conditions and treatments provided for the utterance-related conditions) associated with similar individuals having similar utterance-related conditions as the subject; (iv) provides the health information associated with the similar individuals to a machine learning model to train the machine learning model; and (v) provides, subsequent to the training of the machine learning model, the one or more utterance-related features to the machine learning model to determine one or more treatments for the individual.

First claim

What is claimed is:

1. A system for providing model-based treatment recommendation via individual-specific machine learning models, the system comprising:
one or more processors configured by machine-readable instructions to:
obtain an audio recording of an individual;
determine, from the audio recording, one or more utterance-related features of the individual, the one or more utterance-related features corresponding to characteristics of the individual's utterances in the audio recording;
perform one or more queries based on the one or more utterance-related features to obtain health information associated with similar individuals having similar utterance-related conditions as the subject, the health information indicating utterance-related conditions of the similar individuals and treatments provided to the similar individuals respectively for the utterance-related conditions;
provide the health information associated with the similar individuals to a machine learning model to train the machine learning model; and
provide, subsequent to the training of the machine learning model, the one or more utterance-related features to the machine learning model to determine one or more treatments for the individual.

2. The system of claim 1 , wherein the one or more processors are further configured to:
extract, from the audio recording, a set of utterance-related features of the individual, each utterance-related feature of the set of utterance-related features corresponding to one or more characteristics of the individual's utterances in the audio recording;
perform pattern recognition on the set of utterance-related features to determine which of features of the set of utterance-related features have abnormalities;
determine the one or more utterance-related features by identifying the one or more utterance-related features as features having one or more abnormalities based on the pattern recognition; and
perform the one or more queries (i) based on the one or more utterance-related features and (ii) without reliance on one or more other utterance-related features of the set of utterance-related features to obtain the health information associated with the similar individuals.

3. The system of claim 1 , wherein the one or more processors are further configured to:
for each utterance-related feature of the one or more utterance-related features, determine a classification based on demographic information associated with the individual; and
perform the one or more queries based on the classifications of the one or more utterance-related features to obtain the health information associated with the similar individuals.

4. The system of claim 1 , wherein the one or more processors are further configured to:
extract, from the audio recording, a set of utterance-related features of the individual, each utterance-related feature of the set of utterance-related features corresponding to one or more characteristics of the individual's utterances in the audio recording;
perform first pattern recognition on the set of utterance-related features based on a first pattern recognition scheme to determine which of features of the set of utterance-related features have abnormalities;
perform second pattern recognition on the set of utterance-related features based on a second pattern recognition scheme to determine which of features of the set of utterance-related features have abnormalities;
determine the one or more utterance-related features by identifying a first subset of utterance-related features of the individual as features having one or more abnormalities based on the first pattern recognition, the first subset comprising one or more utterance-related features and other utterance-related features, each utterance-related feature of the first subset corresponding to one or more characteristics of the individual's utterances in the audio recording;
perform the one or more queries (i) based on the first subset of utterance-related features to obtain the health information associated with the similar individuals and (ii) based on the second subset of utterance-related features to obtain other health information associated with other similar individuals having similar utterance-related conditions as the subject, the other health information indicating other utterance-related conditions of the other similar individuals and other treatments provided to the other similar individuals respectively for the other utterance-related conditions;
provide the other health information associated with the other similar individuals to a second machine learning model to train the second machine learning model;
provide, subsequent to the training of the second machine learning model, the second subset of utterance-related features to the second machine learning model to determine one or more other treatments for the individual; and
select a set of treatments for the individual from the one or more treatments and the one or more other treatments.

5. The system of claim 4 , wherein the first pattern recognition scheme is related to at least one of speech waveform recognition, acoustic waveform recognition, speech synthesis recognition, or phonetic sound pronunciation waveform recognition, and wherein the second pattern recognition scheme is related to at least a different one of the speech waveform recognition, the acoustic waveform recognition, the speech synthesis recognition, or the phonetic sound pronunciation waveform recognition.

6. A method for providing model-based treatment recommendation via individual-specific machine learning models, the method being implemented by one or more processors configured by machine readable instructions, the method comprising:
obtaining an audio recording of an individual;
determining, from the audio recording, one or more utterance-related features of the individual, the one or more utterance-related features corresponding to characteristics of the individual's utterances in the audio recording;
performing a query based on the one or more utterance-related features to obtain health information associated with similar individuals having similar utterance-related conditions as the subject, the health information indicating utterance-related conditions of the similar individuals and treatments provided to the similar individuals respectively for the utterance-related conditions;
providing the health information associated with the similar individuals to a machine learning model to train the machine learning model; and
providing, subsequent to the training of the machine learning model, the one or more utterance-related features to the machine learning model to determine one or more treatments for the individual.

7. The method of claim 6 , wherein the method further comprises:
extracting, from the audio recording, a set of utterance-related features of the individual, each utterance-related feature of the set of utterance-related features corresponding to one or more characteristics of the individual's utterances in the audio recording;
performing pattern recognition on the set of utterance-related features to determine which of features of the set of utterance-related features have abnormalities;
determining the one or more utterance-related features by identifying the one or more utterance-related features as features having one or more abnormalities based on the pattern recognition; and
performing the one or more queries (i) based on the one or more utterance-related features and (ii) without reliance on one or more other utterance-related features of the set of utterance-related features to obtain the health information associated with the similar individuals.

8. The method of claim 6 , wherein the method further comprises

Assignees

• Koninklijke Philips Nv

Inventors

• Kempanna Vinutha
• Hariharan Srinivas
• Reddy Siripurapu Mahesh
• Yadalam Kiran Kumar

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G16H20/00

  ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance · CPC title