US-12619667-B2 - Systems and methods for a language model-based customized search platform

US12619667B2US 12619667 B2US12619667 B2US 12619667B2US-12619667-B2

Abstract

Embodiments described herein provide systems and methods for a customized search platform that provides users control and transparency in their searches. The system may use a ranker and parser to utilize input data and contextual information to identify search applications, sort the search applications, and present search results via user-engageable elements. The system may also use input from a user to personalize and update search results based on a user's interaction with user-engageable elements.

Inventors

Bryan McCann
Richard Socher

Assignees

SuSea, Inc.

Dates

Publication Date: 20260505
Application Date: 20240214

Claims (20)

1 . A method of performing an Internet search for generating a search-based response to a user query, the method comprising: receiving, via a communication interface of a search server, a user query; generating, by a neural network based language model trained with past search data and implemented at the search server, an augmented search query based on the user query and a context input; generating, by the neural network based language model, a ranked list comprising a first Internet data source and a second Internet data source that are relevant to the augmented search query for performing an Internet search; generating, by the neural network based language model, a first slot indicating a first volume of search results from the first Internet data source and a second slot indicating a second volume of search results from the second Internet data source, wherein the first volume or the second volume indicate how many search results from the first or the second Internet data source respectively to obtain for presenting to a user; obtaining a first set of search results by performing a first search within the first Internet data source and a second set of search results by performing a second search within the second Internet data source based on the augmented search query, respectively; pulling first content from the first set of search results subject to the first slot and second content from the second set of search results subject to the second slot; causing a display, at a user interface, of a first user-engageable panel displaying the first Internet content with a first indication of the first data source and a second user-engageable panel displaying the second content with a second indication of the second Internet data source.
2 . The method of claim 1 , wherein the augmented search query comprises one or more new words having a contextual relationship with the user query.
3 . The method of claim 1 , wherein the obtaining the first set of search results and the second set of search results comprises: transmitting, via a first search application programming interface (API) integrated at the search server, a first search input customized from the augmented search query to the first Internet data source; and transmitting, via a second search API integrated at the search server, a second search input customized from the augmented search query to the second Internet data source.
4 . The method of claim 1 , further comprising: generating, by a multi-modal neural network based language model, one or more multimodal elements that incorporate the first content or the second content; and presenting, via the user interface, the one or more multimodal elements.
5 . The method of claim 1 , wherein the user query takes a form of a natural language question.
6 . The method of claim 1 , wherein the context input comprises one or more of: user profile information of the user; a conversation history; a conversation context that leads to the user query; user configured preferences or dislikes of one or more data sources; and user past activities approving or disapproving a search result from a specific data source.
7 . The method of claim 1 , wherein the first user-engageable panel and the second user-engageable panel are presented in a ranked order according to the ranked list.
8 . The method of claim 1 , further comprising: receiving, via a first search API or a second search API, additional context information relating to the user query from the first or the second Internet data source; and determining which portion of the input query corresponds to a first search input to the first search API or a second search input to the second search API.
9 . A search system of performing an Internet search for generating a search-based response in response to a search query, the system comprising: a communication interface that receives a user query; a memory storing a plurality of processor-executable instructions and a neural network based language model trained using past search data; a processor coupled to the memory and executing the plurality of processor-executable instructions to perform operations comprising: generating, by the neural network based language model implemented at a search server, an augmented search query based on the user query and a context input; generating, by the neural network based language model, a ranked list comprising a first Internet data source and a second Internet data source that are relevant to the augmented search query for performing an Internet search; generating, by the neural network based language model, a first slot indicating a first volume of search results from the first Internet data source and a second slot indicating a second volume of search results from the second Internet data source, wherein the first volume or the second volume indicate how many search results from the first or the second Internet data source respectively to obtain for presenting to a user; obtaining a first set of search results by performing a first search within the first Internet data source and a second set of search results by performing a second search within the second Internet data source based on the augmented search query, respectively; pulling first content from the first set of search results subject to the first slot and second content from the second set of search results subject to the second slot; causing a display, at a user interface, of a first user-engageable panel displaying the first content with a first indication of the first Internet data source and a second user-engageable panel displaying the second Internet data source with a second indication of the second data source.
10 . The system of claim 9 , wherein the augmented search query comprises one or more new words having a contextual relationship with the user query.
11 . The system of claim 9 , wherein the operation of obtaining the first set of search results and the second set of search results comprises: transmitting, via a first search application programming interface (API) integrated at the search server, a first search input customized from the augmented search query to the first Internet data source; and transmitting, via a second search API integrated at the search server, a second search input customized from the augmented search query to the second Internet data source.
12 . The system of claim 9 , wherein the operations further comprise: generating, by a multi-modal neural network based language model, one or more multimodal elements that incorporate the first content or the second content; and presenting, via the user interface, the one or more multimodal elements.
13 . The system of claim 9 , wherein the user query takes a form of a natural language question.
14 . The system of claim 9 , wherein the context input comprises one or more of: user profile information of the user; a conversation history; a conversation context that leads to the user query; user configured preferences or dislikes of one or more data sources; and user past activities approving or disapproving a search result from a specific data source.
15 . The system of claim 9 , wherein the first user-engageable panel and the second user-engageable panel are presented in a ranked order according to the ranked list.
16 . The system of claim 9 , wherein the operations further comprise: receiving, via a first search API or a second search API, additional context information relating to the user query from the first or the second Internet data source; and determining which portion of the input query corresponds to the first search input to the first search API or the second search input to the second search API.
17 . A non-transitory computer readable medium storing instructions thereon, that when executed by a computing device cause the computing device to perform operations to performing an Internet search for generating a search-based response in response to a search query, the operations comprising: receiving, via a communication interface of a search server, a user query; generating, by a neural network based language model trained with past search data and implemented at the search server, an augmented search query based on the user query and a context input; generating, by the neural network based language model, a ranked list comprising a first Internet data source and a second Internet data source that are relevant to the augmented search query for performing an Internet search; generating, by the neural network based language model, a first slot indicating a first volume of search results from the first Internet data source and a second slot indicating a second volume of search results from the second Internet data source, wherein the first volume or the second volume indicate how many search results from the first or the second Internet data source respectively to obtain for presenting to a user; obtaining a first set of search results by performing a first search within the first Internet data source and a second set of search results by performing a second search within the second Internet data source based on the augmented search query, respectively; pulling first content from the first set of search results subject to the first slot and second content from the second set of search results subject to the second slot; causing a display, at a user interface, of a first user-engageable panel displaying the first content with a first indication of the first data source and a second user-engageable panel displaying the second content with a second indication of the second data source.
18 . The medium of claim 17 , wherein the augmented search query comprises one or more new words having a contextual relationship with the user query.
19 . The medium of claim 17 , wherein the operation of obtaining the first set of search results and the second set of search results comprises: transmitting, via a first search application programming interface (API) integrated at the search server, a first search input customized from the augmented search query to the first Internet data source; and transmitting, via a second search API integrated at the server, a second search input customized from the augmented search query to the second Internet data source.
20 . The medium of claim 17 , wherein the operations further comprise: generating, by a multi-modal neural network based language model, one or more multimodal elements that incorporate the first content or the second content; and presenting, via the user interface, the one or more multimodal elements.

Description

CROSS REFERENCE(S) The instant application is a nonprovisional of and claims priority under 35 U.S.C. 119 to U.S. provisional application 63/484,991, filed Feb. 14, 2023. This instant application is a continuation-in-part of and claims priority under 35 U.S.C. 120 to U.S. non-provisional application Ser. No. 17/981,102, filed Nov. 4, 2022, which in turn claims priority under 35 U.S.C. 119 to U.S. provisional application No. 63/277,091, filed Nov. 8, 2021. All of the above applications are expressly incorporated by reference herein in their entirety. TECHNICAL FIELD The embodiments relate generally to search engines, and more specifically to systems and methods for a customized search platform. BACKGROUND Search engines allow a user to provide a search query and return search results in response. Search sites such as Google.com, Bing.com, and/or the like usually provide a list of search results to a user from all sorts of data sources. For example, these existing search engines usually crawl web data to collect search results that are relevant to a search query. However, a user has little control or transparency on how or where the search engines conduct their search and what kind of search results they are going to get. Therefore, there is a need for a customized search platform that provides users both control and transparency with regard to the searches they perform. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a simplified diagram illustrating a data flow between entities during a search, according to one embodiment described herein. FIG. 2 is a simplified diagram illustrating a computing device implementing the search described in FIG. 1, according to one embodiment described herein. FIG. 3 is a simplified block diagram of a networked system suitable for implementing the search framework described in FIGS. 1-2 and other embodiments described herein. FIG. 4A is a simplified block diagram of a networked system suitable for implementing the customized search platform framework described in FIGS. 1-3 and other embodiments described herein. FIG. 4B is a simplified block diagram of the ranker shown in FIG. 4A, as described with respect to FIG. 4A. FIG. 4C is a simplified block diagram of the parser shown in FIG. 4A, as described with respect to FIG. 4A. FIG. 5 is an example logic flow diagram illustrating a method of search based on the framework shown in FIGS. 1-4A, according to some embodiments described herein. FIG. 6 is an example logic flow diagram illustrating a method of customized search based on the framework shown in FIGS. 1-4A, according to some embodiments described herein. FIG. 7 is an example diagram illustrating an example architecture of a large language model (LLM) based agent for generating a response to a user query, according to according to some embodiments described herein. FIG. 8 is an example diagram illustrating an example architecture of a large language model (LLM) based agent for generating a response to a user query or a prompt for further generation, according to according to some embodiments described herein. FIG. 9 is a simplified block diagram of an example search interface implementing the customized search platform framework described in FIGS. 5-6 and other embodiments described herein. FIGS. 10A-10K are exemplary search interfaces implementing the customized search platform framework described in FIGS. 5-6 and other embodiments described herein. Embodiments of the disclosure and their advantages are best understood by referring to the detailed description that follows. It should be appreciated that like reference numerals are used to identify like elements illustrated in one or more of the figures, wherein showings therein are for purposes of illustrating embodiments of the disclosure and not for purposes of limiting the same. DETAILED DESCRIPTION As used herein, the term “network” may comprise any hardware or software-based framework that includes any artificial intelligence network or system, neural network or system and/or any training or learning models implemented thereon or therewith. As used herein, the term “module” may comprise hardware or software-based framework that performs one or more functions. In some embodiments, the module may be implemented on one or more neural networks. As used herein, the term “Large Language Model” (LLM) may refer to a neural network based deep learning system designed to understand and generate human languages. An LLM may adopt a Transformer architecture that often entails a significant amount of parameters (neural network weights) and computational complexity. For example, LLM such as Generative Pre-trained Transformer (GPT) 3 has 175 billion parameters, Text-to-Text Transfer Transformers (T5) has around 11 billion parameters. The present application generally relates to search engines, and more specifically to systems and methods for a customized search platform. Search engines allow a user to provide a search query and return searc