Classifying retrieved context data for a relativistic response
Methods and systems for managing operation of a system are disclosed. A prompt may be obtained that indicates that relativistic response is to be provided by an inference model. A retrieval process may be performed based on the prompt to obtain context data. The context data may be classified to obtain context data groupings and corresponding meaning descriptions. The context data groupings may be classified based on characteristics of portions of the context data. Ingest data sets may be obtained using the prompt and the context data groupings. The ingest data sets may be provided to the inference model to obtain corresponding responses. The responses may be compared using the meaning descriptions to obtain the relativistic response for use in providing computer-implemented services.
1 . A method for managing operation of a system, the method comprising:
based on a determination that a relativistic response for a prompt submitted for processing by a generative trained machine learning model is to be provided:
performing a retrieval process for the prompt to obtain a plurality of portions of context data from at least one designated data source;
classifying at least the plurality of portions of context data to obtain context data groupings and corresponding meaning descriptions for the context data groupings;
obtaining ingest data sets using the prompt and the context data groupings;
obtaining at least one non-relativistic response using the ingest data sets and the generative trained machine learning model;
obtaining the relativistic response using the at least one non-relativistic response and the corresponding meaning descriptions; and
providing computer-implemented services using the relativistic response,
wherein prior to performing the retrieval process, the method further comprises:
analyzing at least the prompt for indication of desirability of the relativistic response; and
in a first instance of the analyzing where the indication is that the relativistic response is desired: making the determination.
2 . The method of claim 1 , wherein classifying the at least the plurality of portions of context data comprises:
identifying a highest ranked portion of the plurality of portions of the context data;
identifying characteristics of the highest ranked portion; and
clustering the plurality of portions of the context data based on the characteristics to obtain the context data groupings.
3 . The method of claim 2 , wherein during the retrieval process the plurality of portions of the context data are ranked with respect to relevancy to the prompt, and the identifying of the highest ranked portion is performed using the rankings of the plurality of portions of the context data with respect to relevancy to the prompt.
4 . The method of claim 2 , wherein identifying the characteristics of the highest ranked portion comprises at least one selected from a list of identification processes consisting of:
identifying existing tags associated with the highest ranked portion, each of the existing tags ascribing at least one of the characteristics; and
analyzing information context of the highest ranked portion to obtain the characteristics.
5 . The method of claim 2 , wherein each of the context data groupings comprises a portion of the plurality of portions of the context data, and each portion of the plurality of portions of the context data is a member of only one of the context data groupings.
6 . The method of claim 2 , wherein the corresponding meaning descriptions indicate unique meanings for each of the context data groupings, and the corresponding meaning descriptions being based on the characteristics and the context data groupings.
7 . The method of claim 1 , wherein obtaining the ingest data sets comprises:
adding a copy of the prompt to each of the ingest data sets; and
adding members of different context data groupings to the ingest data sets so that each ingest data set has non-duplicative members of the context data groupings with respect to other ingest data sets.
8 . The method of claim 1 , obtaining the at least one non-relativistic response comprises:
separately submitting each of the ingest data sets to the generative trained machine learning model as input data to obtain the at least one non-relativistic response.
9 . The method of claim 1 , wherein obtaining the relativistic response comprises:
obtaining at least one comparative statement between at least two of the at least one non-relativistic response using the corresponding meaning descriptions; and
generating the relativistic response using the at least one comparative statement and the at least two of the at least one non-relativistic response.
10 . The method of claim 1 , further comprising and prior to performing the retrieval process:
in a second instance of the analyzing where the indication is that the relativistic response is not desired: providing second computer implemented services using a second non-relativistic response generated using the generative trained machine learning model.
11 . A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for managing operation of a system, the operations comprising:
based on a determination that a relativistic response for a prompt submitted for processing by a generative trained machine learning model is to be provided:
performing a retrieval process for the prompt to obtain a plurality of portions of context data from at least one designated data source;
classifying at least the plurality of portions of context data to obtain context data groupings and corresponding meaning descriptions for the context data groupings;
obtaining ingest data sets using the prompt and the context data groupings;
obtaining at least one non-relativistic response using the ingest data sets and the generative trained machine learning model;
obtaining the relativistic response using the at least one non-relativistic response and the corresponding meaning descriptions; and
providing computer-implemented services using the relativistic response,
wherein prior to performing the retrieval process, the operations further comprise:
analyzing at least the prompt for indication of desirability of the relativistic response; and
in a first instance of the analyzing where the indication is that the relativistic response is desired: making the determination.
12 . The non-transitory machine-readable medium of claim 11 , wherein classifying the at least the plurality of portions of context data comprises:
identifying a highest ranked portion of the plurality of portions of the context data;
identifying characteristics of the highest ranked portion; and
clustering the plurality of portions of the context data based on the characteristics to obtain the context data groupings.
13 . The non-transitory machine-readable medium of claim 12 , wherein during the retrieval process the plurality of portions of the context data are ranked with respect to relevancy to the prompt, and the identifying of the highest ranked portion is performed using the rankings of the plurality of portions of the context data with respect to relevancy to the prompt.
14 . The non-transitory machine-readable medium of claim 12 , wherein identifying the characteristics of the highest ranked portion comprises at least one selected from a list of identification processes consisting of:
identifying existing tags associated with the highest ranked portion, each of the existing tags ascribing at least one of the characteristics; and
analyzing information context of the highest ranked portion to obtain the characteristics.
15 . The non-transitory machine-readable medium of claim 12 , wherein each of the context data groupings comprises a portion of the plurality of portions of the context data, and each portion of the plurality of portions of the context data is a member of only one of the context data groupings.
16 . A data processing system, comprising:
a processor; and
a memory coupled to the processor to store instructions, which when executed by the processor, cause operations for managing operation of a system to be performed, the operations comprising:
based on a determination that a relativistic response for a prompt submitted for processing by a generative trained machine learning model is to be provided:
performing a retrieval process for the prompt to obtain a plurality of portions of context data from at least one designated data source;
classifying at least the plurality of portions of context data to obtain context data groupings and corresponding meaning descriptions for the context data groupings;
obtaining ingest data sets using the prompt and the context data groupings;
obtaining at least one non-relativistic response using the ingest data sets and the generative trained machine learning model;
obtaining the relativistic response using the at least one non-relativistic response and the corresponding meaning descriptions; and
providing computer-implemented services using the relativistic response,
wherein prior to performing the retrieval process, the operations further comprise:
analyzing at least the prompt for indication of desirability of the relativistic response; and
in a first instance of the analyzing where the indication is that the relativistic response is desired: making the determination.
17 . The data processing system of claim 16 , wherein classifying the at least the plurality of portions of context data comprises:
identifying a highest ranked portion of the plurality of portions of the context data;
identifying characteristics of the highest ranked portion; and
clustering the plurality of portions of the context data based on the characteristics to obtain the context data groupings.
18 . The data processing system of claim 17 , wherein during the retrieval process the plurality of portions of the context data are ranked with respect to relevancy to the prompt, and the identifying of the highest ranked portion is performed using the rankings of the plurality of portions of the context data with respect to relevancy to the prompt.
19 . The data processing system of claim 17 , wherein identifying the characteristics of the highest ranked portion comprises at least one selected from a list of identification processes consisting of:
identifying existing tags associated with the highest ranked portion, each of the existing tags ascribing at least one of the characteristics; and
analyzing information context of the highest ranked portion to obtain the characteristics.
20 . The data processing system of claim 17 , wherein each of the context data groupings comprises a portion of the plurality of portions of the context data, and each portion of the plurality of portions of the context data is a member of only one of the context data groupings.