Schema-based machine-learning model task deduction

Patent metadata

Field	Value
Publication number	US-11599357-B2
Application number	US-202016778554-A
Country	US
Kind code	B2
Filing date	Jan 31, 2020
Priority date	Jan 31, 2020
Publication date	Mar 7, 2023
Grant date	Mar 7, 2023

Abstract

A machine-learning model task deduction method, system, and computer program product include extracting data schema of a machine-learning model and analyzing the data schema to determine an intended task of the machine-learning model.

First claim

What is claimed is:

1. A computer-implemented machine-learning model task deduction method for discovering a utility of a data schema for a machine-learning model, the method comprising:
extracting data schema of a machine-learning model;
analyzing the data schema for classification of an intended task of the machine-learning model;
updating a documentation of the machine-learning model including meta-data and author notes to explicitly identify the intended task;
creating an index of the intended task corresponding to the machine-learning models using the documentation; and
performing a search for an automated discovery of models via the index,
wherein the data schema comprises:
input data; and
output data,
wherein the intended task is determined from the input data and the output data,

wherein the intended task is classified by one or more of:
an application of predetermined rules;
machine-learning; and
creating a schema of the data schema.

2. The method of claim 1 , wherein the task is within a domain including:
vision;
audio; and
natural language.

3. The method of claim 1 , wherein the intended task is selected from a group consisting of:
regression;
classification; and
clustering.

4. The method of claim 1 , embodied in a cloud-computing environment.

5. A computer program product comprising a non-transitory computer readable medium, the computer program product for discovering a utility of a data schema for a machine-learning model comprising a computer-readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform:
extracting data schema of a machine-learning model;
analyzing the data schema for classification of an intended task of the machine-learning model;
updating a documentation of the machine-learning model including meta-data and author notes to explicitly identify the intended task;
creating an index of the intended task corresponding to the machine-learning models using the documentation; and
performing a search for an automated discovery of models via the index,
wherein the data schema comprises:
input data; and
output data,
wherein the intended task is determined from the input data and the output data,

wherein the intended task is classified by one or more of:
an application of predetermined rules;
machine-learning; and
creating a schema of the data schema.

6. The computer program product of claim 5 , wherein the task is within a domain including:
vision;
audio; and
natural language.

7. The computer program product of claim 5 , wherein the intended task is selected from a group consisting of:
regression;
classification; and
clustering.

8. A machine-learning model task deduction system for discovering a utility of a data schema for a machine-learning model, the system comprising:
a processor; and
a memory, the memory storing instructions to cause the processor to perform:
obtaining the data schema of the machine-learning model;
analyzing the data schema for classification of an intended task of the machine-learning model;
updating a documentation of the machine-learning model including meta-data and author notes to explicitly identify the intended task;
creating an index of the intended task corresponding to the machine-learning models using the documentation; and
performing a search for an automated discovery of models via the index,

wherein the data schema comprises:
input data; and
output data,
wherein the intended task is determined from the input data and the output data,

wherein the intended task is classified by one or more of:
an application of predetermined rules;
machine-learning; and
creating a schema of the data schema.

Assignees

• IBM

Inventors

• Braz Alan
• Hirzel Martin
• Shinnar Avraham Ever
• Tsay Jason
• Mummert Todd

Classifications

• G06F8/73 Primary

  Program documentation · CPC title

• G06F8/77

  Software metrics · CPC title

• G06F18/285

  Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/906

  Clustering; Classification · CPC title