# US 12,009,071 B2 · System and/or method for determining service codes from electronic signals and/or states using machine learning

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

- Number: US 12,009,071 B2
- Document ID: US12009071B2
- Status: Granted
- Kind: B2
- Application number: 17/206,021
- Filed: 2021-03-18
- Published: 2022-05-26
- Granted: 2024-06-11
- Earliest claimed: 2020-11-20
- Assignees: AKASA, Inc.
- Applicants: AKASA, Inc.
- Inventors: Byung-Hak Kim; Hariraam Varun Ganapathi
- Web page: https://patentgrep.com/patent/US12009071B2

## CPC

- G06F 40/216: using statistical methods
- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/30: Semantic analysis
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- 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

Disclosed are a system, method and apparatus to generate service codes based, at least in part, on electronic documents. In an embodiment, tokens may be embedded in an electronic document based, at least in part, on a linguistic analysis of the electronic document. Likelihoods of applicability of service codes to the electronic document may be determined based, at least in part, on the embedding of tokens.

## Claims

### Claim 1

A method comprising:
determining an embedding of first tokens based, at least in part, on a linguistic analysis of an electronic document, the electronic document comprising one or more notes relating to a diagnosis and/or treatment created by a medical clinician, at least some of the first tokens to express elements of human anatomy, conditions of elements of human anatomy, or a combination thereof; and
determining likelihoods of applicability of service codes, the service codes identifying clinical services to be performed for medical treatment on a patient, to the electronic document based, at least in part, on the embedding of tokens.

### Claim 2 (depends on 1)

The method of claim 1, wherein determining the embedding of the first tokens further comprises:
partitioning sentences expressed in the electronic document into components including words, partial words and/or punctuation; and
associating the first tokens in a vocabulary of the first tokens with at least some of the components.

### Claim 3 (depends on 2)

The method of claim 2, wherein:
the embedding of first tokens comprises an array of context values associating the first tokens with the service codes identifying clinical services to be performed for medical treatment on the patient; and
determining the embedding of the first tokens further comprises determining a linguistic context of at least some of the first tokens associated with the service codes.

### Claim 4 (depends on 3)

The method of claim 3, wherein determining the linguistic context of at least some of the associated first tokens further comprises applying a bidirectional encoder representations from transformers (BERT).

### Claim 5 (depends on 4)

The method of claim 4, and further comprising training the BERT using jargon, abbreviations, syntax, grammar and/or of text in a medical clinical service domain.

### Claim 6 (depends on 4)

The method of claim 4, and further comprising applying the BERT according to a linguistic domain specific to a medical and/or clinical service.

### Claim 7 (depends on 1)

The method of claim 1, wherein the embedding of first tokens comprises context values associated with individual tokens in a vocabulary of tokens and wherein determining the likelihoods of applicability of service codes to the electronic document further comprises:
applying an attention model to the context values.

### Claim 8 (depends on 7)

The method of claim 7, wherein applying the attention model to the context values further comprises, for computation of a likelihood of applicability of at least at least one of the service codes, computing a dot product of an array of attention coefficients and an array of at least some of the context values associated with the individual tokens.

### Claim 9 (depends on 1)

The method of claim 1, wherein determining likelihoods of applicability service codes to the electronic document further comprises:
applying a self-attention mapping to the embedding of the first tokens based, at least in part, on query, key and value components.

### Claim 10 (depends on 9)

The method of claim 9, wherein determining likelihoods of applicability service codes to the electronic document further comprises:
expressing interrelationships between and/or among code titles, the code titles to identify the service codes, in an embedding of second tokens; and
applying an attention mapping to the embedding of first tokens, the attention mapping being based, at least in part, on the embedding of the second tokens.

### Claim 11

A computing device comprising:
one or more processors to:
determine an embedding of tokens based, at least in part, on a linguistic analysis of an electronic document, the electronic document comprising one or more notes relating to a diagnosis and/or treatment created by a medical clinician, at least some of the tokens to express elements of human anatomy, conditions of elements of human anatomy, or a combination thereof; and
determine likelihoods of applicability of service codes, the service codes identifying clinical services to be performed for medical treatment on a patient, to the electronic document based, at least in part, on the embedding of tokens.

### Claim 12 (depends on 11)

The computing device of claim 11, wherein the one or more processors are further to:
partition sentences expressed in the electronic document into components including words, partial words and/or punctuation; and
associate tokens in a vocabulary of tokens with at least some of the components.

### Claim 13 (depends on 12)

The computing device of claim 12, wherein the one or more processors are to determine the embedding of tokens based on a determination of a linguistic context of at least some of the associated tokens.

### Claim 14 (depends on 13)

The computing device of claim 13, wherein determination of the linguistic context of at least some of the associated tokens further comprises application of a bidirectional encoder representations from transformers (BERT).

### Claim 15 (depends on 14)

The computing device of claim 14, wherein the one or more processors are further to train the BERT using jargon, abbreviations, syntax, grammar and/or of text in a medical clinical service domain.

### Claim 16 (depends on 14)

The computing device of claim 14, wherein the one or more processors are further to apply the BERT according to a linguistic domain specific to a medical and/or clinical service.

### Claim 17 (depends on 11)

The computing device of claim 11, wherein the embedding of tokens comprises context values associated with individual tokens in a vocabulary of tokens and wherein determination of the likelihoods of applicability of the service codes to the electronic document further comprises:
application of an attention model to the context values.

### Claim 18 (depends on 17)

The computing device of claim 17, wherein application of the attention model to the context values further comprises, for computation of a likelihood of applicability of at least at least one of the service codes, computation of a dot product of an array of attention coefficients and an array of at least some of the context values associated with the individual tokens.

### Claim 19

An article comprising:
a non-transitory storage medium comprising computer-readable instructions stored thereon which are executable by one or more processors of a computing device to:
determine an embedding of first tokens based, at least in part, on a linguistic analysis of an electronic document, the electronic document comprising one or more notes relating to a diagnosis and/or treatment created by a medical clinician, at least some of the first tokens to express elements of human anatomy, conditions of elements of human anatomy, or a combination thereof; and
determine likelihoods of applicability of service codes, the service codes identifying clinical services to be performed for medical treatment on a patient, to the electronic document based, at least in part, on the embedding of tokens.