Creating Diversity in Artificial Intelligence and Machine Learning
AssigneeSony Interactive Entertainment Inc.
InventorAlbhy Galuten
Life of this patent
Estimating the term…
Abstract
The present invention concern systems and methods for maintaining diversity in AI and ML environments through the cooperation of various AI and ML systems such that they optimize for social and cultural diversity. The examination of behavior, infrastructure, and governance, mimicking of genetic biodiversity, and application of the foregoing to machine reasoning mitigates the tendency of systems to find optimized or single best solutions. AI and ML environments may thus derive multiple diverse solutions that contribute to richer ecosystems in which human beings may function and thrive.
Claims21
A system for maintenance of diversity of machine learning datasets, the system comprising:
- memory that maintains one or more different datasets concerning infrastructure, governance, and behaviors;
- a communication interface that communicates over a communication network to receive a selection of a data set from among the different datasets;
- a processing device that executes a software engine maintained in a non-transitory storage medium, the software engine executed by the processing device to:
- identify one or more groupings including at least one niche grouping within the selected data set and one or more relationships between the groupings, wherein two or more of the groupings share data in common, and wherein data regarding the groupings is maintained separately for each of the groupings;
- introduce cross-pollination of different groupings into the selected data set, wherein the data set is updated further based on the cross-pollination;
- track the relationships between the groupings after the cross-pollination;
- monitor the niche grouping for changes in niche population after the cross-pollination;
- and
- generate a report on diversity that includes information regarding health of the grouping and ability of the grouping to sustain within a corresponding ecosystem.
A method for maintenance of diversity of machine learning datasets, the method comprising:
- maintaining in memory one or more different datasets concerning infrastructure, governance, and behaviors;
- communicating via a communication interface over a communication network to receive a selection of a data set from among the different datasets;
- identifying one or more groupings including at least one niche grouping within the selected data set and one or more relationships between the groupings, wherein two or more of the groupings share data in common, and wherein data regarding the groupings is maintained separately for each of the groupings;
- introducing cross-pollination of different groupings into the selected data set, wherein the data set is updated further based on the cross-pollination;
- tracking the relationships between the groupings after the cross-pollination;
- monitoring the niche grouping for changes in niche population after the cross-pollination;
- and
- generating a report on diversity that includes information regarding health of the grouping and ability of the grouping to sustain within a corresponding ecosystem.
A non-transitory, computer-readable storage medium, having embodied thereon a program executable by a processor to perform a method for maintenance of diversity of machine learning datasets, the method comprising:
- maintaining in memory one or more different datasets concerning infrastructure, governance, and behaviors;
- communicating via a communication interface over a communication network to receive a selection of a data set from among the different datasets;
- identifying one or more groupings including at least one niche grouping within the selected data set and one or more relationships between the groupings, wherein two or more of the groupings share data in common, and wherein data regarding the groupings is maintained separately for each of the groupings;
- introducing cross-pollination of different groupings into the selected data set, wherein the data set is updated further based on the cross-pollination;
- tracking the relationships between the groupings after the cross-pollination;
- monitoring the niche grouping for changes in niche population after the cross-pollination;
- and
- generating a report on diversity that includes information regarding health of the grouping and ability of the grouping to sustain within a corresponding ecosystem.
Dependencies are inferred from the claim text.