Systems and methods for using a generative model to generate an output based on a set of diverse inputs

Patent metadata

Field	Value
Publication number	US-2025335750-A1
Application number	US-202418651418-A
Country	US
Kind code	A1
Filing date	Apr 30, 2024
Priority date	Apr 30, 2024
Publication date	Oct 30, 2025
Grant date	—

Abstract

A method for compromising between competing ideas includes generating a group of embedded inputs by embedding each input of a group of inputs in an embedding space. The method also includes weighting each embedded input of the group of embedded inputs based on one or more parameters. The method further includes calculating a weighted centroid based on weighting each embedded input. The method still also includes generating, via a generative model, an output based on the weighted centroid.

First claim

What is claimed is:

1 . A method for compromising between competing ideas, comprising:
generating a group of embedded inputs by embedding each input of a group of inputs in an embedding space; weighting each embedded input of the group of embedded inputs based on one or more parameters; calculating a weighted centroid based on weighting each embedded input; and generating, via a generative model, an output based on the weighted centroid.

2 . The method of claim 1 , wherein each input of the group of inputs is a solution to a problem.

3 . The method of claim 2 , wherein the output is a novel solution to the problem.

4 . The method of claim 1 , wherein the one or more parameters include a distance from an initial centroid and/or one or more factor values.

5 . The method of claim 4 wherein:
each factor value of the one or more factor values is associated with a different factor of a group of factors; and
the group of factors include at least one of cost, equity, diversity, or sustainability.

6 . The method of claim 4 , wherein the initial centroid is based on the group of embedded inputs, prior to weighting each embedded input.

7 . The method of claim 1 , wherein the output is different than each input of the group of inputs.

8 . An apparatus for compromising between competing ideas, comprising:
one or more processors; and one or more memories coupled with the one or more processors and storing instructions operable, when executed by the one or more processors, to cause the apparatus to:
generate a group of embedded inputs by embedding each input of a group of inputs in an embedding space;
weight each embedded input of the group of embedded inputs based on one or more parameters;
calculate a weighted centroid based on weighting each embedded input; and
generate, via a generative model, an output based on the weighted centroid.

9 . The apparatus of claim 8 , wherein each input of the group of inputs is a solution to a problem.

10 . The apparatus of claim 9 , wherein the output is a novel solution to the problem.

11 . The apparatus of claim 8 , wherein the one or more parameters include a distance from an initial centroid and/or one or more factor values.

12 . The apparatus of claim 11 wherein:
each factor value of the one or more factor values is associated with a different factor of a group of factors; and
the group of factors include at least one of cost, equity, diversity, or sustainability.

13 . The apparatus of claim 11 , wherein the initial centroid is based on the group of embedded inputs, prior to weighting each embedded input.

14 . The apparatus of claim 8 , wherein the output is different than each input of the group of inputs.

15 . A non-transitory computer-readable medium having program code recorded thereon for compromising between competing ideas, the program code executed by a processor and comprising:
program code to generate a group of embedded inputs by embedding each input of a group of inputs in an embedding space; program code to weight each embedded input of the group of embedded inputs based on one or more parameters; program code to calculate a weighted centroid based on weighting each embedded input; and program code to generate, via a generative model, an output based on the weighted centroid.

16 . The non-transitory computer-readable medium of claim 15 , wherein:
each input of the group of inputs is a solution to a problem; and the output is a novel solution to the problem.

17 . The non-transitory computer-readable medium of claim 15 , wherein the one or more parameters include a distance from an initial centroid and/or one or more factor values.

18 . The non-transitory computer-readable medium of claim 17 wherein:
each factor value of the one or more factor values is associated with a different factor of a group of factors; and
the group of factors include at least one of cost, equity, diversity, or sustainability.

19 . The non-transitory computer-readable medium of claim 17 , wherein the initial centroid is based on the group of embedded inputs, prior to weighting each embedded input.

20 . The non-transitory computer-readable medium of claim 15 , wherein the output is different than each input of the group of inputs.

Assignees

• Toyota Res Inst Inc
• Toyota Motor Co Ltd

Inventors

• Chen Francine R
• Lau Tatiana

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N3/0475 Primary

  Generative networks · CPC title