ATTRIBUTION OF GENERATIVE MODEL OUTPUTS

Number: 20250021820
Published: 2025-01-16
Filed: 2024-02-22
Assignee: DISNEY ENTERPRISES, INC.
Inventors: DOGGETT; Erika Varis et al.
CPC: G06Q50/184; G06N3/0475; G06N3/0455; G06F16/55; G06N20/00; G06N3/0464; G06N3/0895; G06Q10/10; G06Q30/0276; G06Q10/04
Verdict: Set aside generative-model output attribution/governance, business/legal
Source: Google Patents · FreePatentsOnline

Abstract

The present invention sets forth a technique for analyzing a generative output of a generative model. The technique includes determining a first latent representation of the generative output and a plurality of latent representations of a plurality of data samples associated with the generative model. The technique also includes computing a plurality of similarities between the first latent representation and the plurality of latent representations. In response to determining that a first similarity that is included in the plurality of similarities and computed between the first latent representation and a second latent representation included in the plurality of latent representations exceeds a threshold, the technique includes causing output to be generated that indicates a high similarity between the generative output and a first data sample that is included in the plurality of data samples and corresponds to the second latent representation.

Background

BACKGROUND Field of the Various Embodiments

Embodiments of the present disclosure relate generally to machine learning and generative models and, more specifically, to attribution of generative model outputs. Description of the Related Art

Generative models refer to deep neural networks and/or other types of machine learning models that are trained to generate new instances of data. For example, a generative model could be trained on a training dataset of images of cats. During training, the generative model “learns” the visual attributes of the various cats depicted in the images. These learned visual attributes could then be used by the generative model to produce new images of cats that are not found in the training dataset.

A generative model is typically trained on a large dataset of existing content, which can include copyrighted content, offensive content, and/or other types of restricted content. When the trained generative model is subsequently used to generate new samples of content, the generated samples can be identical to, resemble, and/or otherwise be similar to certain content items within the restricted content. To avoid potential issues associated with violating copyrights, content policies, and/or other issues associated with generating and/or using restricted content, these types of generated samples should be identified before the generated samples are outputted, distributed, and/or published. This typically involves attributing a give

Claims

1. A computer-implemented method for analyzing a generative output of a generative model, the method comprising: determining a first latent representation of the generative output and a plurality of latent representations of a plurality of data samples associated with the generative model; computing a plurality of similarities between the first latent representation and the plurality of latent representations; determining that a first similarity that is included in the plurality of similarities and computed between the first latent representation and a second latent representation included in the plurality of latent representations exceeds a threshold; and in response to determining that the first similarity exceeds the threshold, causing output to be generated that indicates a high similarity between the generative output and a first data sample that is included in the plurality of data samples and corresponds to the second latent representation. || 11. One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of: determining a first latent representation of a generative output of a generative model and a plurality of latent representations of a plurality of data samples associated with the generative model; computing a plurality of similarities between the first latent representation and the plurality of latent representations; determining that a first similarity that is included in the plurality of similarities and computed between the first latent representation and a second latent representation included in the plurality of latent representations exceeds a threshold; and in response to determining that the first similarity exceeds the threshold, causing output to be generated that indicates a high similarity between the generative output and a first data sample that is included in the plurality of data samples and corresponds to the second latent representation. || 20. A system, comprising: one or more memories that store instructions, and one or more processors that are coupled to the one or more memories and, when executing the instructions, are configured to perform the steps of: generating a latent representation of an output of a machine learning model and a plurality of latent representations of a plurality of data samples associated with the machine learning model; computing a plurality of similarities between the latent representation and the plurality of latent representations; determining that a first similarity that is included in the plurality of similarities and computed between the latent representation and a first data sample included in the plurality of data samples exceeds a threshold; and in response to determining that the first similarity exceeds the threshold, causing additional output indicating a high similarity between the output and the first data sample to be generated.