Guiding machine learning models and related components

Patent metadata

Field	Value
Publication number	US-11537932-B2
Application number	US-201715840315-A
Country	US
Kind code	B2
Filing date	Dec 13, 2017
Priority date	Dec 13, 2017
Publication date	Dec 27, 2022
Grant date	Dec 27, 2022

Abstract

Techniques facilitating guiding machine learning models and related components are provided. In one example, a computer-implemented method comprises identifying, by a device operatively coupled to a processor, a set of models, wherein the set of models includes respective model components; determining, by the device, one or more model relations among the respective model components, wherein the one or more model relations respectively comprise a vector of component relations between respective pairwise ones of the model components; and suggesting, by the device, a subset of the set of models based on a mapping of the component relations.

First claim

What is claimed is:

1. A system comprising:
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a suggestion component that receives a user query for suggested models, wherein the user query comprises model-related data;
an identification component that identifies models, wherein the models respectively comprise model components;
a learning component trains a cognitive exploration system comprising a neural network to determine model relations among the model components and the model-related data;
a determination component that determines, using the cognitive exploration system comprising the neural network, the model relations among the model components and the model-related data, wherein the model relations comprise vectors of component relations between respective pairwise ones of the model components and the model-related data; and
the suggestion component responds to the user query with the suggested models comprising a subset of the models selected, using the cognitive exploration system comprising the neural network, based on the vectors of component relations, wherein the subset of the models are selected to be most dissimilar to the model-related data according to a tolerance parameter indicative of degree of similarity.

2. The system of claim 1 , wherein the computer executable components further comprise:
a distance component that computes distances between the pairwise ones of the model components and the model-related data.

3. The system of claim 2 , wherein the subset of the models are selected based further on the distances as represented in the vectors of component relations.

4. The system of claim 3 , wherein the subset of the models are selected based further on a comparison of the distances to the tolerance parameter.

5. The system of claim 1 , wherein the subset of the models are selected to be similar to the model-related data according to another tolerance parameter indicative of degree of similarity with respect to a first criterion, and are selected to be most dissimilar to the model-related data according to the tolerance parameter indicative of degree of similarity with respect to a second criterion.

6. The system of claim 1 , wherein the model components comprise at least one of model configurations, model program code, model training data, model feedback, deployment data, or parent model information.

7. A computer-implemented method comprising:
receiving, by a device operatively coupled to a processor, a user query for suggested models, wherein the user query comprises model-related data;
identifying, by the device, models, wherein the models respectively comprise model components;
training, by the device, a cognitive exploration system comprising a neural network to determine model relations among the model components and the model-related data;
determining, by the device, using the cognitive exploration system comprising the neural network, the model relations among the model components and the model-related data, wherein the model relations comprise respective vectors of component relations between respective pairwise ones of the model components and the model-related data; and
responding, by the device, to the user query with the suggested models comprising a subset of the models selected, using the cognitive exploration system comprising the neural network, based on the vectors of component relations, wherein the subset of the models are selected to have a high dissimilarity to the model-related data according to a tolerance parameter indicative of degree of similarity.

8. The computer-implemented method of claim 7 , wherein the subset of the models are selected to have a high similarity to the model-related data according to another tolerance parameter indicative of degree of similarity with respect to a first criterion, and are selected to have the high dissimilarity to the model-related data according to the tolerance parameter indicative of degree of similarity with respect to a second criterion.

9. The computer-implemented method of claim 7 , further comprising:
computing, by the device, distances between the pairwise ones of the model components and the model-related data.

10. The computer-implemented method of claim 9 , further comprising selecting, by the device, using the cognitive exploration system comprising the neural network, the subset of the models based on the distances as represented in the vectors of component relations.

11. The computer-implemented method of claim 10 , wherein the selecting is further based on a comparison of the distances to the tolerance parameter.

12. The computer-implemented method of claim 7 , wherein the model components comprise at least one of model configurations, model program code, model training data, model feedback, deployment data, or parent model information.

13. A computer program product for providing guidance in machine learning models, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:
receive a user query for suggested models, wherein the user query comprises model-related data;
identify models, wherein the models respectively comprise model components;
train a cognitive exploration system comprising a neural network to determine model relations among the model components and the model-related data;
determine, using the cognitive exploration system comprising the neural network, the model relations among the model components and the model-related data, wherein the model relations comprise respective vectors of component relations between respective pairwise ones of the model components and the model-related data; and
respond to the user query with the suggested models comprising a subset of the models selected, using the cognitive exploration system comprising the neural network, based on the vectors of component relations, wherein the subset of the models are selected to be of greater dissimilarity to the model-related data according to a tolerance parameter indicative of degree of similarity.

14. The computer program product of claim 13 , wherein the subset of the models are selected to be of greater similarity to the model-related data according to another tolerance parameter indicative of degree of similarity with respect to a first criterion, and are selected to be of the greater dissimilarity to the model-related data according to the tolerance parameter indicative of degree of similarity with respect to a second criterion.

15. The computer program product of claim 13 , wherein the program instructions further cause the processor to:
compute distances between the pairwise ones of the model components and the model-related data.

16. The computer program product of claim 15 , wherein the program instructions further cause the processor to:
select, using the cognitive exploration system comprising the neural network, the subset of the models based on the distances as represented in the vectors of component relations.

17. The computer program product of claim 13 , wherein the model components comprise at least one of model configurations, model program code, model training data, model feedback, deployment data, or parent model information.

18. A system comprising:
a memory that stores computer executable components; and
a processor that executes the computer e

Assignees

• IBM

Inventors

• Bobroff Norman
• Braz Alan
• Hirzel Martin
• Mummert Todd
• Westerink Peter

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N20/20 Primary

  Ensemble learning · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00 Primary

  Machine learning · CPC title