Interactive search and generation system for social identity of objects data using large language model(s)
Machine learning based processing systems and techniques are described. In some examples, a machine learning based processing system analyzes text-based input to extract a plurality of natural language elements from the text-based input. The text-based input is associated with an object. The machine learning based processing system generates a prompt from at least a subset of the plurality of natural language elements. The machine learning based processing system analyzes the prompt by using a trained machine learning model to generate a response. The response is responsive to the prompt. The machine learning based processing system analyzes the response to extract a plurality of media content elements from the response. The plurality of media content elements corresponds to different aspects of the object.
1 . A method for scored searching, the method comprising:
analyzing a text-based input to extract a plurality of natural language elements from the text-based input, wherein the text-based input is associated with a physical object;
generating a prompt from at least a subset of the plurality of natural language elements, wherein the prompt represents a search query for a search;
analyzing the prompt using a trained machine learning model to generate a response, wherein the response is responsive to the prompt and represents a search result of the search;
analyzing the response to extract a plurality of media content elements from the response, wherein the plurality of media content elements corresponds to different aspects of the physical object; and
analyzing the plurality of media content elements to generate a score, wherein the score is associated with at least one of an accuracy of the search result or a responsiveness of the search result to the search query.
2 . The method of claim 1 , further comprising receiving the text-based input with a user interface.
3 . The method of claim 1 , further comprising:
querying at least one data structure using a data structure query to retrieve contextual data, wherein the data structure query is based on the text-based input; and
modifying the prompt using the contextual data before analyzing the prompt using the trained machine learning model.
4 . The method of claim 1 , wherein the plurality of natural language elements includes a plurality tokens.
5 . The method of claim 1 , wherein the score is based on cross-referencing of the plurality of media content elements with one or more data sources.
6 . The method of claim 1 , further comprising:
adjusting the score based on a comparison between the plurality of media content elements and the different aspects of the physical object.
7 . The method of claim 1 , further comprising:
receiving feedback, wherein the feedback is based on a user input; and
refining the plurality of media content elements based on the feedback.
8 . The method of claim 7 , further comprising:
analyzing the plurality of media content elements after refining the plurality of media content elements to identify additional media content elements that correspond to additional aspects of the physical object; and
associating the additional media content elements by providing references to the additional aspects of the physical object.
9 . The method of claim 1 , further comprising:
receiving a voice clip; and
interpreting the voice clip using a speech-to-text algorithm to generate the text-based input.
10 . The method of claim 1 , wherein the plurality of media content elements includes one or more Social Identify of Objects (SIO) data elements.
11 . The method of claim 1 , further comprising:
analyzing the plurality of media content elements to identify a shared attribute of at least a subset of the plurality of media content elements; and
searching a data structure for the shared attribute to retrieve one or more additional media content elements from the data structure.
12 . The method of claim 1 , wherein the different aspects of the physical object include at least one of people, places, physical properties, origination, emotions, cultures, or events.
13 . The method of claim 1 , further comprising:
filtering a subset of the different aspects of the physical object from the response based on the prompt.
14 . The method of claim 1 , wherein the response has at least one of a natural language format or a table format.
15 . The method of claim 1 , further comprising:
receiving feedback associated with the plurality of media content elements; and
updating the trained machine learning model based on the feedback to improve an accuracy of the trained machine learning model.
16 . A system for scored searching, the system comprising:
a memory that stores instructions; and
a processor that executes the instructions, wherein execution of the instructions by the processor causes the processor to:
analyze a text-based input to extract a plurality of natural language elements from the text-based input, wherein the text-based input is associated with a physical object;
generate a prompt from at least a subset of the plurality of natural language elements, wherein the prompt represents a search query for a search;
analyze the prompt using a trained machine learning model to generate a response, wherein the response is responsive to the prompt and represents a search result of the search;
analyze the response to extract a plurality of media content elements from the response, wherein the plurality of media content elements corresponds to different aspects of the physical object; and
analyze the plurality of media content elements to generate a score, wherein the score is associated with at least one of an accuracy of the search result or a responsiveness of the search result to the search query.
17 . The system of claim 16 , wherein the execution of the instructions by the processor causes the processor to:
query at least one data structure using a data structure query to retrieve contextual data, wherein the data structure query is based on the text-based input; and
modify the prompt using the contextual data before analyzing the prompt using the trained machine learning model.
18 . The system of claim 16 , wherein the execution of the instructions by the processor causes the processor to receive the text-based input with a user interface.
19 . The system of claim 16 , wherein the score is based on cross-referencing of the plurality of media content elements with one or more data sources.
20 . The system of claim 16 , wherein the execution of the instructions by the processor causes the processor to:
adjust the score based on a comparison between the plurality of media content elements and the different aspects of the physical object.