Managing inference models using structured representations for summarized inferences

Patent metadata

Field	Value
Publication number	US-2025371314-A1
Application number	US-202418678628-A
Country	US
Kind code	A1
Filing date	May 30, 2024
Priority date	May 30, 2024
Publication date	Dec 4, 2025
Grant date	—

Abstract

Methods and systems for managing a generative inference model are disclosed. Using the generative inference model and ingest data, an inference may be obtained. To determine whether the inference is acceptable (e.g., for downstream use, in view of reference works), a summarization data package for the inference may be obtained using at least a summarization model. The summarization data package for the inference may be used to levels of similarity of the inference with respect to reference works. Acceptability of the inference may be determined based on the levels of similarity and a similarity threshold. If the inference is unacceptable, performance of an action set may be initiated to manage an impact of similarities between the inference and at least one of the reference works.

First claim

What is claimed is:

1 . A method for managing a generative inference model, the method comprising:
obtaining an inference generated using the generative inference model and ingest data; obtaining a summarization data package for the inference using at least a summarization model; obtaining, using at least in part the summarization data package for the inference, levels of similarity of the inference with respect to reference works; making a determination regarding whether the inference is acceptable based on the levels of similarity and a similarity threshold; and in a first instance of the determination where the inference is acceptable:
providing the inference to a downstream consumer as a computer-implemented service; and
in a second instance of the determination where the inference is unacceptable:
initiating performance of an action set to manage an impact of similarities between the inference and at least one of the reference works.

2 . The method of claim 1 , wherein the summarization model comprises a foundation model adapted to extract a subset of information from a source data object.

3 . The method of claim 2 , wherein the foundation model is further adapted to limit a quantity of information added to the summarization data package for the inference.

4 . The method of claim 3 , wherein the foundation model is adapted to use a summary schema to generate the summarization data package for the inference, the summary schema discriminating the subset of the information from other information from the source data object.

5 . The method of claim 4 , wherein the summary schema discriminates conceptual information from the source data object from contextual information from the source data object.

6 . The method of claim 1 , wherein the generative inference model is trained to generate human interpretable text when prompted using the ingest data, the human interpretable text being responsive to a request indicated by the ingest data.

7 . The method of claim 6 , further comprising:
obtaining a structured representation for the inference based on the summarization data package for the inference using a structured representation schema, wherein the structured representation for the inference is used during the obtaining of the of the levels of similarity as a basis of comparison for the inference to the reference works.

8 . The method of claim 7 , wherein the structured representation schema is adapted to facilitate identification of at least one of:
a location indicated by the inference; a character indicated by the inference; an object indicated by the inference; and a law of nature indicated by the inference.

9 . The method of claim 1 , wherein obtaining the summarization data package for the inference comprises:
obtaining an objective, the objective indicating at least one constraint for adding information to the summarization data package for the inference; and prompting, using at least the objective, the summarization model to generate the summarization data package for the inference.

10 . The method of claim 1 , further comprising:
prior to obtaining the levels of similarity:
for a portion of the reference works:
obtaining a summarization data package for the portion of the reference works using at least the summarization model; and
obtaining a structured representation for the portion of the reference works based on the summarization data package for the portion of the reference works using a structured representation schema.

11 . The method of claim 7 , wherein the structured representation for the inference comprises a graph-structured data model that specifies relationships between elements of the summarization data package for the inference, the graph-structured data model comprising nodes and edges, and the edges being based on the relationships between the elements associated with the edges.

12 . The method of claim 11 , wherein obtaining the levels of similarity comprises:
performing a sub-graph analysis of the structured representation for the inference with respect to portions of structured representations for the reference works to identify whether a portion of the structured representation for the inference substantially matches one of the portions of the structured representations for the reference works.

13 . The method of claim 1 , wherein the levels of similarity indicate likelihoods that the inference plagiarizes the reference works.

14 . The method of claim 1 , wherein the action set comprises obtaining a description of the similarities between the inference and the at least one of the reference works.

15 . The method of claim 1 , wherein the action set comprises preventing provision of the inference to the downstream consumer.

16 . The method of claim 1 , wherein the action set comprises at least one action that, when performed, modifies operation and/or use of the generative inference model to reduce a likelihood that a future inference generated using the generative inference model and the ingest data plagiarizes the reference works.

17 . A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing a generative inference model, the operations comprising:
obtaining an inference generated using the generative inference model and ingest data; obtaining a summarization data package for the inference using at least a summarization model; obtaining, using at least in part the summarization data package for the inference, levels of similarity of the inference with respect to reference works; making a determination regarding whether the inference is acceptable based on the levels of similarity and a similarity threshold; and in a first instance of the determination where the inference is acceptable:
providing the inference to a downstream consumer as a computer-implemented service; and
in a second instance of the determination where the inference is unacceptable:
initiating performance of an action set to manage an impact of similarities between the inference and at least one of the reference works.

18 . The non-transitory machine-readable medium of claim 17 , wherein the summarization model comprises a foundation model adapted to extract a subset of information from a source data object.

19 . A data processing system, comprising:
a processor; and a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations for managing a generative inference model, the operations comprising:
obtaining an inference generated using the generative inference model and ingest data;
obtaining a summarization data package for the inference using at least a summarization model;
obtaining, using at least in part the summarization data package for the inference, levels of similarity of the inference with respect to reference works;
making a determination regarding whether the inference is acceptable based on the levels of similarity and a similarity threshold; and
in a first instance of the determination where the inference is acceptable:
providing the inference to a downstream consumer as a computer-implemented service; and

in a second instance of the determination where the inference is unacceptable:
initiating performance of an action set to manage an impact of similarities between the inference and at least one of the reference works.

20

Assignees

• Dell Products Lp

Inventors

• Ezrielev Ofir
• Kushnir Tomer
• Savir Amihai

Classifications

• G06N20/00

  Machine learning · CPC title

• G06N3/0455 Primary

  Auto-encoder networks; Encoder-decoder networks · CPC title