# US 2021/0081841 A1 · Visually creating and monitoring machine learning models

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

- Number: US 2021/0081841 A1
- Document ID: US20210081841A1
- Status: Published application
- Kind: A1
- Application number: 17/017,594
- Filed: 2020-09-10
- Published: 2021-03-18
- Earliest claimed: 2019-09-12
- Assignees: Viani Systems, Inc.
- Applicants: Viani Systems, Inc.
- Inventors: Vishal Inder SIKKA; Daniel James AMELANG; Kevin Frederick DUNNELL
- Web page: https://patentgrep.com/patent/US20210081841A1

## CPC

- G06N 3/105: {Shells for specifying net layout}
- G06N 20/00: Machine learning
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/048: Activation functions
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 3/0985: Hyperparameter optimisation; Meta-learning; Learning-to-learn
- G06N 5/04: Inference or reasoning models
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

One embodiment of the present invention sets forth a technique for creating a machine learning model. The technique includes generating a user interface comprising one or more components for visually generating the machine learning model. The technique also includes modifying source code specifying a plurality of mathematical expressions that define the machine learning model based on user input received through the user interface. The technique further includes compiling the source code into compiled code that, when executed, causes one or more parameters of the machine learning model to be learned during training of the machine learning model.

## Claims

### Claim 1

A method for creating a machine learning model, comprising:
generating a user interface comprising one or more components for visually generating the machine learning model;
modifying source code specifying a plurality of mathematical expressions that define the machine learning model based on user input received through the user interface; and
compiling the source code into compiled code that, when executed, causes one or more parameters of the machine learning model to be learned during training of the machine learning model.

### Claim 2 (depends on 1)

The method of claim 1, further comprising modifying a visual representation of the machine learning model in the user interface based on the user input.

### Claim 3 (depends on 2)

The method of claim 2, wherein the visual representation comprises one or more layers of the machine learning model, one or more neurons in the one or more layers, one or more features inputted into the machine learning model, and one or more outputs of the machine learning model.

### Claim 4 (depends on 3)

The method of claim 3, wherein the visual representation further comprises a layer type associated with the one or more layers, an activation function associated with the one or more layers, and a model type of the machine learning model.

### Claim 5 (depends on 1)

The method of claim 1, further comprising outputting, in the user interface, one or more additional components for managing one or more objectives associated with the machine learning model.

### Claim 6 (depends on 5)

The method of claim 5, wherein the one or more objectives comprise at least one of a project schedule, a label to be predicted, a threshold for a performance metric associated with the label, and a source of training data for the machine learning model.

### Claim 7 (depends on 1)

The method of claim 1, further comprising outputting, in the user interface, one or more additional components for managing one or more experiments associated with the machine learning model.

### Claim 8 (depends on 7)

The method of claim 7, wherein the one or more additional components comprise an experiment version, a dataset version, a model version of the machine learning model, and an experiment status.

### Claim 9 (depends on 1)

The method of claim 1, further comprising outputting, in the user interface, one or more additional components for interacting with a training result of training the machine learning model.

### Claim 10 (depends on 9)

The method of claim 9, wherein the one or more additional components comprise at least one of a precision-recall curve, a confusion matrix, a training dataset for the machine learning model, and a filter associated with the training dataset.

### Claim 11 (depends on 1)

The method of claim 1, wherein compiling the source code into the compiled code comprises:
generating an abstract syntax tree (AST) representation of the source code;
generating the compiled code based on the AST representation; and
determining that the one or more parameters in the machine learning model is to be learned based on a structure of the source code.

### Claim 12 (depends on 1)

The method of claim 1, wherein the one or more components comprise a component for specifying at least a portion of the source code for defining the machine learning model.

### Claim 13 (depends on 1)

The method of claim 1, further comprising upon generating the compiled code, incrementing one or more versions associated with the machine learning model and an experiment comprising the machine learning model.

### Claim 14

A non-transitory computer readable medium storing instructions that, when executed by a processor, cause the processor to perform the steps of:
generating a user interface comprising one or more components for visually generating a machine learning model;
modifying source code specifying a plurality of mathematical expressions that define the machine learning model based on user input received through the user interface;
modifying a visual representation of the machine learning model in the user interface based on the user input; and
compiling the source code into compiled code that, when executed, causes one or more parameters of the machine learning model to be learned during training of the machine learning model.

### Claim 15 (depends on 14)

The non-transitory computer readable medium of claim 14, wherein the steps further comprise outputting, in the user interface, one or more additional components for managing (i) one or more objectives associated with the machine learning model and (ii) one or more experiments associated with the machine learning model.

### Claim 16 (depends on 15)

The non-transitory computer readable medium of claim 15, wherein the one or more additional components comprise an experiment version, a dataset version, a model version of the machine learning model, and an experiment status.

### Claim 17 (depends on 14)

The non-transitory computer readable medium of claim 14, wherein the steps further comprise outputting, in the user interface, one or more additional components for interacting with a training result of training the machine learning model.

### Claim 18 (depends on 17)

The non-transitory computer readable medium of claim 17, wherein the one or more additional components comprise at least one of a precision-recall curve, a confusion matrix, a training dataset for the machine learning model, and a filter associated with the training dataset.

### Claim 19 (depends on 14)

The non-transitory computer readable medium of claim 14, wherein the visual representation comprises one or more layers of the machine learning model, one or more neurons in the one or more layers, one or more features inputted into the machine learning model, one or more outputs of the machine learning model, a layer type associated with the one or more layers, an activation function associated with the one or more layers, or a model type of the machine learning model.

### Claim 20

A system, comprising:
a memory that stores instructions, and
a processor that is coupled to the memory and, when executing the instructions, is configured to:
generate a user interface comprising one or more components for visually generating a machine learning model;
modify source code specifying a plurality of mathematical expressions that define the machine learning model based on user input received through the user interface;
modify a visual representation of the machine learning model in the user interface based on the user input;
compile the source code into compiled code that, when executed, causes one or more parameters of the machine learning model to be learned during training of the machine learning model; and
upon generating the compiled code, increment one or more versions associated with the machine learning model and an experiment comprising the machine learning model.
