Selective presentation of rich experiences in search

By recording user browsing history and analyzing their interests, combining rich clip experience and confidence scores, the problem that search engines in the prior art is difficult to provide personalized search results, and more relevant and useful search results are achieved.

CN113574522BActive Publication Date: 2024-12-06MICROSOFT TECHNOLOGY LICENSING LLC
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202080020946.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-03-14
Filing Date
2020-01-31
Publication Date
2024-12-06
Estimated Expiration
2040-01-31

AI Technical Summary

Technical Problem

Existing search engines find it difficult to provide personalized search results based on the search intentions and interests of different users, resulting in poor user experience.

Method used

By recording and analyzing the user's browsing history, use the rich clip experience associated with interest to present with search results, and determine whether to display the rich clip experience based on the user's relevance and confidence scores to provide personalized search results.

Benefits of technology

Improve the relevance and user experience of search results, ensure that search results are more in line with users' expectations and intentions, and enhance the usefulness of search results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113574522B_ABST
    Figure CN113574522B_ABST
Patent Text Reader

Abstract

A method for selectively providing different types of search engine results to different searchers is provided. Browsing history for each of a plurality of unique identities is recorded. In response to the browsing history for the unique identity satisfying relevance criteria for a rich snippet experience, the unique identity is associated with the rich snippet experience. The rich snippet experience is configured to present featured snippet-specific information on a search results webpage along with other search results. In response to receiving a search query from a unique identity previously associated with the rich snippet experience, the rich snippet experience is presented on the search results webpage. In response to receiving a search query from a different unique identity that was not previously associated with the rich snippet experience, other search results without the rich snippet experience are presented on the search results webpage.
Need to check novelty before this filing date? Find Prior Art

Description

Background Art

[0001] Internet search engines and other search providers may be designed to provide many different search results in response to a search query. Internet search engines may be configured to populate search results web pages with different types of search results. For example, a traditional search results web page may include a list of uniform resource locators (URLs) for related websites. Such URLs themselves do not provide information related to the search, but only provide links to web pages that may contain relevant information. Some search engines may also be configured to present a rich segment experience with other URLs, which is itself a summary of useful information related to the search, rather than just links to other information sources. The rich segment experience may include information related to the search query that is selected and presented in a visually distinct manner relative to other conventional URL search results. The rich segment experience may make it easier for users to quickly digest relevant information related to the search query without having to explore other URLs. Summary of the invention

[0002] This summary is provided to introduce a selection of concepts that are further described in the detailed description below in a simplified form. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all of the disadvantages noted in any part of this disclosure.

[0003] A method for selectively providing different types of search engine results to different searchers is provided. Browsing history for each of a plurality of unique identities is recorded. In response to the browsing history for the unique identity satisfying relevance criteria for a rich snippet experience, the unique identity is associated with the rich snippet experience. The rich snippet experience is configured to present selected snippet-specific information on a search results webpage along with other search results. In response to receiving a search query from a unique identity previously associated with the rich snippet experience, the rich snippet experience is presented on the search results webpage. In response to receiving a search query from a different unique identity that was not previously associated with the rich snippet experience, other search results are presented on the search results webpage without the rich snippet experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Figure 1 An example search engine computing system is schematically illustrated.

[0005] Figure 2 is a flow chart depicting aspects of an example method for recording browsing history information for a unique identity.

[0006] Figure 3is a flow chart depicting aspects of an example method for selectively providing different types of search engine results to different searchers.

[0007] Figure 4A-4F An example search results web page is shown with different search results that selectively include a rich snippet experience based on different unique identities' browsing histories.

[0008] Figure 5 An example computing system is schematically illustrated. DETAILED DESCRIPTION

[0009] In some cases, different searchers may perform the same search query, but with different intent or different interpretations of the search query. Prior to the search strategy disclosed herein, different searchers may be provided with search results that include the same rich snippet experience associated with the search query by the search engine. For example, different searchers may provide the search query "Rams" to the search engine. The first searcher may be researching the Los Angeles Rams football team. The second searcher may be researching Dodge brand Ram trucks. The third searcher may be researching male sheep. Depending on the specific rich snippet experience that the search engine associates with the search query, at least some of the different searchers will not find the search results useful if the same rich snippet experience is provided to all searchers.

[0010] Thus, the present disclosure relates to a method for determining a searcher's interest in particular experiences / snippets and using these interests as a factor in deciding whether to present a particular rich snippet experience in response to a search query. By selectively providing rich snippet experiences based on the searcher's interests, the searcher's intent can be anticipated and more useful search results tailored to the particular searcher can be provided.

[0011] Figure 1An example computing architecture is shown that enables different types of search engine results to be provided to different searchers. The search engine computing system 100 can be configured to provide search functionality to multiple client computers 102 (e.g., 102A, 102B, 102C). For example, the search engine computing system 100 can enable the client computers 102 to perform searches on any suitable search domain (e.g., (multiple) websites, (multiple) encyclopedias, (multiple) databases, social networking platforms, financial / medical / scientific records, etc.) using any suitable query (e.g., natural language keywords, regular expression patterns, Boolean operators for composing multiple queries, etc.). The search engine computing system 100 can be communicatively coupled to each of the multiple client computers 102 via a computer network 104 such as the Internet. The search engine computing system 100 can be configured to provide search functionality to any suitable number of different client computers 102.

[0012] The plurality of client computers 102 may be any suitable type of computing device, such as a mobile phone, a personal computer, a smart assistant speaker device, etc. One or more of the plurality of client computers 102 may optionally include a display subsystem configured to present a web browser 106. The web browser 106 may be configured to present one or more pages, such as a search page configured to allow a user to enter a search query and / or view search results, examples of which are described in more detail below. Figure 4A-4F . Alternatively or additionally, one or more of the plurality of client computers 102 may allow a user to input a search query and / or receive search results in any other suitable manner, for example, one or more of the plurality of client computers 102 may include a microphone and a speaker and may be configured to receive a query via a natural language speech utterance received at the microphone and to output search results via a speech audio output at the speaker. Although the present disclosure may be described in terms of a graphical web browser 106, the methods and processes disclosed herein may be implemented in the context of any other search interface (e.g., a speech-based natural language search interface).

[0013] Different client computers 102 can be associated with different unique identities 108. Unique identities 108 can take any suitable form. In some examples, unique identities 108 can be persistent user identities across different computing devices. For example, a user can perform search queries on a desktop computer, a tablet computer, and a smart phone, and these search queries are all associated with the same unique identity, and all of these search queries can be attributed to the same unique identity. In some such examples, the unique identity can take the form of an anonymous identifier (ANID) anonymously linked to a user account. In other examples, the unique identity 108 can be associated with a specific client computer. For example, the search behavior and / or browsing history of different users sharing the same client computer 102 can be tracked via the same unique identity 108 associated with the client computer 102. In some such examples, the unique identity can take the form of a client computing device identifier (client ID).

[0014] In the depicted example, client computer 102A and client computer 102B are associated with unique identity A, such that search behavior and browsing history tracked on each of client computers 102A and 102B are attributed to unique identity A. For example, both client computer 102A and client computer 102B may be associated with the same user. In addition, client computer 102C is associated with unique identity B, such that search behavior and browsing history tracked on client computer 102C are attributed to unique identity B, which is different from unique identity A.

[0015] The search engine computing system 100 is configured to return one or more search results on a search results web page in response to receiving a search query from a client computer 102 for presentation in a web browser 106. The search engine computing system 100 may be configured to serve "original" queries in the form of literal text input by a user. Alternatively or additionally, the search engine computing system 100 may be configured to serve "normalized" queries in the form of computer-readable descriptions of query content, for example, by processing a computer-readable description indicating query intent (query intent represents the user's question, goal, and / or task indicated by the query), by processing one or more entities in the query, and / or by processing the syntactic structure of the query (e.g., a parse tree for the query). Query normalization may be performed by any suitable computer device(s), for example, by a client computer 102 and / or a search engine computing system 100. Normalized queries may include relevant information content (e.g., relevant intent / entities) of the query while limiting the amount of variability between queries (e.g., different original queries that reformulate the same question may be normalized to the same normalized query).

[0016] The search results may include different types of search results retrieved from different sources. The search engine computing system 100 may be configured to return results from any suitable domain, such as different websites on the Internet or other domains. The search engine computing system 100 may return search results based on data received from one or more other computers. For example, the search engine computing system 100 may collaborate with a search result provider 110 to send a search query to the search result provider 110, and receive search results for the search query, which may include relevant websites, documents, etc. for the desired search domain. The search result provider 110 may be configured to determine relevant search results for the query in any suitable manner, such as by searching in a database, indexing / crawl websites or documents, etc.

[0017] Non-limiting examples of search result items that may be presented on a search results web page may include non-selected search results, such as a list of URLs of related websites. Such URLs themselves do not provide information relevant to the search, but only provide links to web pages that may have relevant information. Alternatively or in addition to non-selected search results, the search results web page may selectively include a rich snippet experience 132. Each rich snippet experience 132 may present a summary of useful information related to the search query, rather than just links to other information sources. Unlike a list of URLs, a rich snippet experience is selected to provide enough information so that the user does not need to explore the search results page in depth, unlike clicking a URL to navigate to a third-party website.

[0018] The rich fragment experience may include information related to the search query that is selected and presented in a visually distinct manner relative to other non-selected URL search results. In some examples, the rich fragment experience 132 may include graphics, images, animations, and / or videos. The rich fragment experience 132 may have visual characteristics that draw attention to the rich fragment experience. For example, the rich fragment experience 132 may be prominently positioned above other non-selected search results on the search results web page. As another example, the rich fragment experience 132 may be positioned on a side panel of the search results web page. In some examples, the rich fragment experience 132 may be formatted differently from other non-selected search results. For example, the rich fragment experience 132 may be contained within a box or may be separated from other search results by some form of border.

[0019] Search engine computing system 100 can be configured to maintain multiple rich snippet experiences 132 in experience database 130. Each different rich snippet experience 132 can relate to a different topic, theme, subject, or other type of information. Experience database 130 can be configured to store any suitable number of different rich snippet experiences 132.

[0020] Search engine computing system 100 may include a ranker 112 configured to rank the relevance of each search result item relative to the search query, whether it is a rich snippet experience or a non-selected search result, such as a URL of a website. Ranker 112 may assign different rankings to different search result items, and may present different search result items on a search result web page based on the rankings. Ranker 112 may be configured to rank different search result items according to any suitable ranking technique. In some implementations, ranker 112 may be configured to rank rich snippet experience 132 separately from other search results.

[0021] In one example, the ranker 112 may be configured to selectively present different rich snippet experiences 132 on the search results web page based on a triggering algorithm 140. The ranker 112 may use the triggering algorithm 140 to determine the relevance of different rich snippet experiences 132 stored in the experience database 130 to a particular search query. The triggering algorithm 140 may include one or more thresholds for triggering the presentation of the rich snippet experience 132. Each rich snippet experience 132 may provide the ranker 112 with a confidence score 134 determined based on the search query.

[0022] Confidence scores can be generated in any suitable manner, for example, using any suitable combination of state-of-the-art and / or future machine learning (ML), artificial intelligence (AI), and / or natural language processing (NLP) technologies. For example, confidence scores can be generated by AI, ML, and / or NLP models based on input data, including any suitable combination of search queries, personalized data for users, and / or data related to candidate rich fragment experiences to be provided to search queries. In some examples, the model can be trained to generate confidence scores for classification tasks compared to "real data" examples of relevant rich fragment experiences, for example, for classifying relevance, predicting user satisfaction, and / or correctly generating relevant rich fragment experiences. For example, a model can be given input data and configured to output a relevance score, a user satisfaction score, and / or select a rich fragment experience. The output of the model can be evaluated according to a loss function (e.g., a loss function that measures the accuracy of the score and / or selection). The model can be configured to output one or more confidence values ​​for the classification task (e.g., confidence values ​​for different classification results of user satisfaction and / or relevance, confidence values ​​for the relevance of different candidate rich fragment experiences, etc.). Similarly, alternatively or additionally, in addition to using a loss function on a classification task, the model can also be trained via reinforcement learning on a reinforcement signal, e.g., feedback on user satisfaction or user-provided relevance of a selected rich snippet. In either case, whether using a loss function or reinforcement learning, the system can be adjusted over time to "punish" incorrect actions by adjusting system parameters so that incorrect actions are less likely to occur in the future given similar inputs, and to "reward" correct actions by adjusting system parameters so that correct actions are more likely to be made in the future. The loss function and / or reinforcement can be configured to reward and / or penalize answers more heavily in proportion to the confidence value. Thus, the confidence value can be adjusted based on the loss function, e.g., to penalize confident but incorrect answers, and / or to reward confident correct answers. Thus, the system can be trained to emit relatively high confidence values ​​when it is likely to be correct, and to emit relatively low confidence values ​​when the correct answer cannot be predicted with confidence.

[0023] In some examples, different rich snippet experiences may use different scoring techniques to determine confidence scores. Ranker 112 may be configured to compare confidence score 134 to a threshold, and if confidence score 134 is greater than the threshold, rich snippet experience 132 may be presented. On the other hand, if the confidence score is less than the threshold, presentation of rich snippet experience 132 may be suppressed.

[0024] In some implementations, the ranker 112 may optionally be configured to position the rich snippet experience on the search results web page based on the confidence score 134. For example, a rich snippet with a higher confidence score may be placed more prominently on the search results page (e.g., at the top of the page above other search results). As another example, a rich snippet with a medium confidence score that is just enough to exceed a threshold for presenting a rich snippet experience may be placed less prominently on the search results web page (e.g., below or to the side of the page).

[0025] In addition, in some examples where multiple rich snippet experiences 132 are triggered, the ranker 112 can be configured to select the triggered rich snippet experience with the highest confidence score 134 for presentation (or more prominent presentation). In some such examples, multiple rich snippet experiences 132 can be presented on the search results web page. In addition, in some examples, different portions of the search results web page can have different priorities, such that the rich snippet experience with the highest confidence score can be positioned at the highest priority position on the search results web page (e.g., at the top of the page). The ranker 112 can traverse the list to position different rich snippet experiences at different priority positions on the search results web page based on the confidence score. The rich snippet experience can be positioned at any suitable position on the search results web page based on the confidence of the rich snippet experience.

[0026] As described above, the searcher's interests can be used as a factor in determining whether to display a particular rich snippet experience in response to a search query. In some examples, the threshold used by the trigger algorithm 140 to determine whether to present a rich snippet experience may depend on whether the searcher has previously shown interest in the rich snippet and has been associated with the rich snippet experience. In some such examples, if the unique identity 108 of the provided search query is associated with a rich snippet experience, the ranker 112 may be configured to use a relaxed trigger threshold. In addition, if the unique identity 108 is not associated with a rich snippet experience, the ranker 112 may be configured to use a strict trigger threshold greater than the relaxed trigger threshold. For example, since the search query for "Ninja" is quite broad, the non-personalized strict trigger threshold that triggers the "eSports" rich snippet experience may be 0.8 or 80%. However, if the unique identity providing the search query was previously associated with the "eSports" rich snippet experience, a relaxed trigger threshold of 0.6 or 60% may trigger the "eSports" rich snippet experience. For any suitable rich snippet experience, the strict trigger threshold and the relaxed trigger threshold can be set to any suitable threshold. In some examples, the trigger threshold may be determined based on the search query. For example, a broader search query may have a lower threshold, while a narrower search query may have a higher threshold.

[0027] Additionally or alternatively, in some examples, confidence score 134 provided by rich snippet experience 132 may be increased or decreased based on whether unique identity 108 is associated with rich snippet experience 132. For example, the confidence score of a rich snippet experience may be increased based on the unique identity associated with the rich snippet experience. By using these different thresholds and / or varying confidence scores based on the searcher's interest / association with the rich snippet experience, a rich snippet experience may be more likely to be triggered to be presented on a search results webpage relative to a rich snippet experience that other searchers have not previously expressed interest in.

[0028] Rich snippet experiences 132 may be associated with one or more unique identities 138 based on relevance criteria 136. In some examples, different rich snippet experiences 132 may have different relevance criteria 136. Rich snippet experiences 132 may use any suitable relevance criteria 136 to determine whether unique identity 108 has expressed interest in rich snippet experience 132 and should therefore be associated with rich snippet experience 132. For example, relevance criteria 136 may be based on natural language processing, machine learning, artificial intelligence, data mining, based on direct one-to-one matching, "fuzzy" matching (e.g., matching with at least a threshold similarity), and / or probabilistic matching. When a unique identity 108 meets relevance criteria 136 for a rich snippet experience 132, unique identity 108 may be associated with rich snippet experience 132 in experience database 130. As a natural extension of this example, a cluster or set of associated unique identities 138 may be formed for each rich snippet experience 132 and stored in experience database 130 along with the rich snippet experience 132.

[0029] In one non-limiting example, each rich snippet experience 132 can be defined according to a set of descriptive keyword tags. In addition, each unique identity 108 can have a set of descriptive keyword tags that are associated with the unique identity based on previous search behavior and / or browsing history 126. Each rich snippet experience 132 can compare its set of keyword tags to the keyword tags of each unique identity. When the keyword tag set of a rich snippet experience 132 matches (or matches within at least a threshold similarity) the keyword tags associated with the unique identity 108, the relevance criteria 136 of the rich snippet experience 132 can be satisfied and the unique identity 108 can be associated with the rich snippet experience 132 in the experience database 130.

[0030] The search engine computing system 100 may include a search behavior / browsing history pipeline 114 that can track the search behavior / browsing history of each unique identity 108 to facilitate determining the correlation between the unique identity and the rich snippet experience. The record pipeline 114 may include a unique identity extractor 116 that is configured to identify the unique identity 108 that provides the search query from the client computer 102 to the search engine computing system 100. The record pipeline 114 may include a search query tagger 118 that is configured to extract keywords from the search query. The search query tagger 118 may perform natural language processing (NLP) techniques on the search query, for example, a part of speech (POS) tagger may be used to obtain keyword tags from the search query.

[0031] The recording pipeline 114 may include a search result tagger 120 configured to obtain keyword tags from search behavior / browsing history. In particular, the search result tagger 120 may be configured to track which search results are selected from the search result web page (e.g., identifying which URLs are clicked). In addition, the search result tagger 120 may be configured to perform NLP techniques on the selected search results. For example, a text snippet may be extracted from the URL, and a POS tagger may be used to obtain keyword tags from the text snippet. The search result tagger 120 may be configured to obtain keywords from a rich fragment experience presented on the search result web page. The search result tagger 120 may be configured to obtain keywords from search results of other granular search verticals such as news search results, image search results, video search results, shopping search results, recipe search results, etc.

[0032] The record pipeline 114 may optionally include a search result filter 122, which may be configured to remove specified search results for keyword tags used to generate the unique identity 108. In some implementations, the search result filter 122 may be configured to filter out selected search results that are considered non-active (e.g., causing abandonment) so as not to be used to generate keyword tags for the unique identity. For example, the search result filter 122 may be configured to filter out selected search results whose dwell time is less than a threshold time. As an example, a searcher may click on a URL that renders a website in a web browser, and then the searcher may quickly navigate back to the search result webpage to select a different search result. In this case, the search result filter 122 may filter out the selected search results. In some examples, the search result filter 122 may be configured to filter out other specified search results that may be considered private or unwanted (specified by the searcher or computing system). As an example, the search result filter 122 may be configured to filter out search results related to adult content. As another example, the search result filter 122 may be configured to filter out search results related to privacy information (e.g., medical information).

[0033] The keyword tags and other search behavior / browsing history information 126 tracked by the recording pipeline 114 for each unique identity 108 may be maintained in the identity database 124. The identity database 124 may be configured to store any suitable number of unique identities 108 and associated browsing histories 126. The identity database 124 may store any suitable information related to the unique identities 108. In some examples, each unique identity 108 may be associated with a different set of keyword tags aggregated for the unique identity 108 based on the search behavior / browsing history in the identity database 124.

[0034] As discussed above, the set of keyword tags associated with each rich snippet experience 132 can be compared to the keyword tags associated with each unique identity 108 to form different clusters of unique identities 138 that are interested in different rich snippet experiences 132. Each unique identity cluster can be associated with a rich snippet experience in the experience database 130. Alternatively or additionally, a different rich snippet experience cluster 128 can be associated with each unique identity 108 in the identity database 124. In either case, the associated information can be provided to the ranker 112, and the ranker 112 can adjust the trigger algorithm 140 - for example, by using a relaxed / strict trigger threshold and / or adjusting the confidence score based on such information. By adjusting the trigger algorithm 140 in this manner, rich snippet experiences in which the searcher has previously expressed interest can be more likely to be presented on the search results webpage relative to other rich snippet experiences in which the searcher has not previously expressed interest. In this way, the search results are more likely to be useful and meet the expectations of the searcher.

[0035] Figure 2 Depicted are aspects of an example method 200 for recording browsing history information for a unique identity. The method 200 may be performed by any suitable computer system, for example, by the search engine computing system 100 and / or the client computer 102. For example, the search engine computing system 100 may instantiate the recording pipeline 114 to perform the method 200.

[0036] At 202, method 200 includes receiving a search query from a client computer. At 204, method 202 includes identifying a unique identity that provided the search query. In some examples, the unique identity may be an ANID associated with a searcher and may be used across multiple different client computers. In some examples, the unique identity may be a client ID associated with a particular client computer.

[0037] At 206, method 200 includes generating keyword tags derived from the search query. Such keyword tags can be included in a set of keyword tags associated with the unique identity that characterizes interests of the unique identity.

[0038] At 208, method 200 includes providing search results based on the search query on a search results webpage presented in a web browser of the client computer.

[0039] At 210, method 200 includes generating keyword tags obtained from the search results. In some implementations, at 212, the method may optionally include generating keyword tags obtained from the selected search results. As an example, the keyword tags may be obtained from a text snippet included in the selected URL. As another example, the keyword tags may be obtained from information about a website selected from the search results web page. In some implementations, at 214, method 200 may optionally include generating keyword tags obtained from a rich snippet experience presented on the search results web page. In some implementations, at 216, method 200 may optionally include generating keyword tags obtained from other vertical search results (e.g., image search, video search, person search, recipe search). Such keyword tags may be included in a keyword tag set associated with a unique identity that characterizes the interests of the unique identity.

[0040] At 218, method 200 may optionally include filtering specified search results to avoid being used to generate keyword tags. As an example, search results that are considered to be non-positive (e.g., resulting in abandonment) can be filtered out to not be used to generate keyword tags for unique identities. Other non-limiting examples include filtering out search results related to adult content and privacy information.

[0041] At 220, method 200 includes associating the unique identity with one or more rich snippet experiences based on a relevance criterion. As an example, if a set of keyword tags associated with a rich snippet experience and a keyword tag associated with the unique identity at least partially match, the relevance criterion is met and the unique identity is associated with the rich snippet experience.

[0042] Method 200 may be performed for each unique identity in a plurality of unique identities to determine the interests of each unique identity and the rich snippet experience associated with each unique identity. Such information may be used to provide different types of search engine results to different searchers, as discussed herein.

[0043] Figure 3 Aspects of an example method 300 for selectively providing different types of search engine results to different searchers are depicted. The method 300 may be performed by any suitable computer system, such as the search engine computing system 100 and / or the client computer 102.

[0044] At 302, method 300 includes receiving a search query from a unique identity. Method 300 may perform a series of method steps in parallel for each of a plurality of rich snippet experiences to determine whether a rich snippet experience should be presented on a search results web page based on the search query. For each rich snippet experience, at 304, method 300 includes determining whether the unique identity is associated with the rich snippet experience. If the unique identity is associated with the rich snippet experience, method 300 moves to 306. Otherwise, the unique identity is not associated with the rich snippet experience, and method 300 moves to 308.

[0045] At 306, method 300 includes determining whether the rich snippet experience is greater than a relaxation trigger threshold. For example, the rich snippet experience can generate a confidence score based on the search query and can compare the confidence score to the relaxation trigger threshold. If the rich snippet experience is greater than the relaxation trigger threshold, method 300 moves to 312. Otherwise, method 300 moves to 310.

[0046] At 308, method 300 includes determining whether the rich snippet experience is greater than a strict trigger threshold, which is greater than a relaxed trigger threshold. For example, the rich snippet experience can generate a confidence score based on the search query and can compare the confidence score to the strict trigger threshold. If the rich snippet experience is greater than the strict trigger threshold, method 300 moves to 312. Otherwise, method 300 moves to 310.

[0047] At 310, method 300 includes presenting other search results on the search results web page without presenting the rich snippet experience. For example, the other search results may include URLs to related websites and / or different, more relevant rich snippet experiences.

[0048] At 312, method 300 includes presenting the rich snippet experience on the search results web page along with other search results. In some examples, the rich snippet experience can optionally be presented on the search results web page without any non-selected search results (e.g., URL links) or any other search results.

[0049] In some implementations, at 314, method 300 can optionally include positioning the rich snippet experience on the search results webpage based on the confidence score of the rich snippet experience. For example, the higher the confidence score of the rich snippet experience, the higher the position of the rich snippet experience on the search results webpage.

[0050] Method steps 304 to 314 may be repeated for each of the plurality of rich snippet experiences to determine whether each rich snippet experience is to be presented on the search results webpage.

[0051] 4A to 4FAn example search result webpage with different search results that selectively include a rich snippet experience based on the search behavior / browsing history of different unique identities is shown. In particular, search results for the same search query performed by unique identity A and unique identity B are shown. Unique identity A is interested in the collectible card game Magic the Gathering, which is determined based on previous search behavior / browsing history for unique identity A. Unique identity B is interested in the Orlando Magic basketball team, which is determined based on previous search behavior / browsing history for unique identity B.

[0052] refer to Figure 4A , search results web page 400 includes a search query 402 for “MAGIC THE GATHERING” performed by unique identity A. Search query 402 may be provided as an input to a triggering algorithm. Each of a plurality of different rich snippet experiences may determine a confidence score based on search query 402. The triggering algorithm may be configured to compare the plurality of confidence scores to different triggering thresholds. In particular, the confidence score of a rich snippet experience associated with unique identity A may be compared to a relaxed triggering threshold, and the confidence score of a rich snippet experience not associated with unique identity A may be compared to a strict triggering threshold.

[0053] The rich snippet experience 404 of "MAGIC: THE GATHERING" is associated with unique identity A because unique identity A is interested in this type of collectible card game experience. Therefore, the confidence score of rich snippet experience 404 is compared to the relaxed trigger threshold. The confidence score of rich snippet experience 404 exceeds the relaxed trigger threshold, which triggers rich snippet experience 404 to be presented on the search results webpage 400 for unique identity A.

[0054] The rich snippet experience 404 includes images of different cards and other selected information related to game rules, articles about the game, game play tournaments, cards for sale, and downloads of digital versions of the game. The selected information can be retrieved from different websites and presented in the rich snippet experience 404. Additionally, other non-selected search results 406 are presented on the search results webpage 400 based on the search query 402 performed by the unique identity A. The other non-selected search results 406 include URLs to different related websites. In some examples, the other non-selected search results 406 can be personalized for the unique identity A. In other examples, the other non-selected search results 406 can be non-personalized, so that the same search results are presented to different unique identities based on the search query 402.

[0055] In addition, "ORLANDO MAGIC BASKETBALL TEAM" ( Figure 4D ) is not relevant to unique identity A because unique identity A has not previously expressed interest in this basketball team (e.g., via previous search and / or browsing activity). Therefore, the confidence score of rich snippet experience 412 is compared to the strict trigger threshold. The confidence score of rich snippet experience 412 does not exceed the strict trigger threshold and therefore is not presented on search results webpage 400 for unique identity A based on search query 402.

[0056] exist Figure 4B In the search result web page 400, the search result web page 400 includes Figure 4A 4. The same search query 402 for “MAGIC THE GATHERING” performed by unique identity A in FIG. 40 is performed by unique identity B instead. The rich snippet experience 404 for “MAGIC: THE GATHERING” is not relevant to unique identity B because unique identity B has not previously demonstrated an interest in this type of collectible card game experience (e.g., via previous search and / or browsing activity). Therefore, the confidence score of the rich snippet experience 404 is compared to the strict trigger threshold. The confidence score of the rich snippet experience 404 exceeds the strict trigger threshold, which triggers the rich snippet experience 404 to be presented on the search results webpage 400 for unique identity B. In addition, based on the search query 402 performed by unique identity B, other non-featured search results 408 are presented on the search results webpage 400.

[0057] In addition, the rich snippet experience 412 for “ORLANDO MAGIC BASKETBALL TEAM” (eg Figure 4D 4) is associated with unique identity B because unique identity B is interested in this basketball team. Therefore, the confidence score of rich snippet experience 412 is compared to the relaxation trigger threshold. The confidence score of rich snippet experience 412 does not exceed the relaxation trigger threshold, and therefore will not be presented on the search results webpage 400 for unique identity B based on search query 402.

[0058] Notice, Figure 4B 4. It is shown that even though the unique identity was previously associated with rich snippet experience 412 for "ORLANDO MAGIC BASKETBALL TEAM", if the search query is specific enough for another rich snippet experience, the unique identity may not be provided with that rich snippet experience (e.g., rich snippet experience 404). Further note that Figure 4A and Figure 4BTogether, it is shown that the same rich snippet experience can be provided to different unique identities with different interests who enter the same search query if the rich snippet experience is specific enough to that search query.

[0059] exist Figure 4C , search results webpage 400 includes a search query 408 for “MAGIC ORLANDO” performed by unique identity A. The confidence score of a rich snippet experience 404 for “MAGIC: THE GATHERING” is compared to a slack trigger threshold because rich snippet experience 404 is associated with unique identity A. The confidence score of rich snippet experience 404 exceeds the slack trigger threshold, which triggers rich snippet experience 404 to be presented on search results webpage 400 for unique identity A. Additionally, other non-featured search results 410 are presented on search results webpage 400 based on search query 408 performed by unique identity A. Other search results 410 include URLs to different relevant websites for search query 408, including MAGIC TOURNAMENTS IN ORLANDO and SHOPS IN ORLANDO SELLING MAGIC CARDS.

[0060] In addition, the "ORLANDO MAGIC BASKETBALL TEAM" (such as Figure 4D 4 , is compared to the strict trigger threshold because rich snippet experience 412 is associated with unique identity A. The confidence score of rich snippet experience 412 does not exceed the strict trigger threshold and is therefore not presented on search result webpage 400 for unique identity A based on search query 408 .

[0061] exist Figure 4D , search results web page 400 includes a search query 408 for “MAGIC ORLANDO” performed by unique identity B. The confidence score of rich snippet experience 412 for “ORLANDO MAGIC BASKETBALL TEAM” is compared to the relaxed trigger threshold because rich snippet experience 412 is associated with identity B. Rich snippet experience 412 has a confidence score that exceeds the relaxed trigger threshold, which triggers rich snippet experience 412 to be presented on search results web page 400 for unique identity B.

[0062] The rich snippet experience 412 includes scores for the Orlando Magic basketball team's currently ongoing game, as well as other curated information related to the basketball team's roster, schedule, standings, statistics, and tickets. The curated information may be retrieved from different websites and presented in the rich snippet experience 412. Additionally, other non-curated search results 414 are presented on the search results webpage 400 based on the search query 408 performed by the unique identity B. The other non-curated search results 414 include URLs to different relevant websites for the search query 416.

[0063] In addition, "MAGIC: THE GATHERING" ( Figure 4C ) is compared to the strict trigger threshold because rich snippet experience 404 is not associated with unique identity B. The confidence score for rich snippet experience 404 does not exceed the strict trigger threshold, and therefore is not presented on search result webpage 400 for unique identity B based on search query 408.

[0064] Notice, Figure 4C and Figure 4D Together, it is shown that, based on the previous association of different unique identities with different rich snippet experiences, different rich snippet experiences can be provided to different unique identities even if those unique identities enter the same search query. Figure 4E , search results web page 400 includes a search query 416 for “ORLANDO MAGIC BASKETBALL” performed by unique identity A. The confidence score 412 of the rich snippet experience for “ORLANDO MAGIC BASKETBALL TEAM” is compared to the strict trigger threshold because rich snippet experience 412 is not associated with unique identity A. The confidence score of rich snippet experience 412 exceeds the strict trigger threshold, which triggers rich snippet experience 412 to be presented on search results web page 400 for unique identity A.

[0065] Additionally, other non-selected search results 414 are presented on the search results webpage 400 based on the search query 416 performed by the unique identity A. The other non-selected search results 414 include URLs to different related websites based on the search query 416.

[0066] In addition, "MAGIC: THE GATHERING" ( Figure 4C ) is compared to the relaxation trigger threshold because the rich snippet experience 404 is associated with the unique identity A. The confidence score 404 of the rich snippet experience does not exceed the relaxation trigger threshold, and therefore is not presented on the search results webpage 400 for the unique identity A based on the search query 416.

[0067] Notice, Figure 4E It is shown that even though the unique identity was previously associated with the rich snippet experience 404 for “MAGIC: THE GATHERING,” the unique identity may not be provided a rich snippet experience (eg, rich snippet experience 412 ) if the search query is specific enough for another rich snippet.

[0068] refer to Figure 4F , search results web page 400 includes a search query 416 for “ORLANDO MAGICBASKETBALL” performed by unique identity B. The confidence score 412 of the rich snippet experience for “ORLAND MAGIC BASKETBALL TEAM” is compared to the relaxed trigger threshold because rich snippet experience 412 is associated with unique identity B. The confidence score of rich snippet experience 412 exceeds the relaxed trigger threshold, which triggers rich snippet experience 412 to be presented on search results web page 400 for unique identity B. Additionally, based on search query 416 performed by unique identity B, other non-featured search results 414 are presented on search results web page 400.

[0069] In addition, "MAGIC:THE GATHERING" ( Figure 4C ) is compared to the strict trigger threshold because rich snippet experience 404 is not associated with unique identity B. The confidence score of rich snippet experience 404 does not exceed the strict trigger threshold and is therefore not presented on search result webpage 400 for unique identity B based on search query 416.

[0070] Notice, Figure 4E and Figure 4F Together, it is shown that the same rich snippet experience can be provided to different unique identities with different interests who enter the same search query if the rich snippet experience is specific enough to that search query.

[0071] Although the concepts described herein are intended to present different search results to different searchers on a search results web page, it should be understood that different search results can take any other suitable form. For example, a rich snippet experience can be a natural language result of speech (e.g., through a personal assistant Alexa / Siri / Cortana).

[0072] The methods and processes described herein may be bound to a computing system of one or more computing devices. In particular, such methods and processes may be implemented as an executable computer application, a network accessible computing service, an application programming interface (API), a library, or a combination of the above and / or other computing resources.

[0073] Figure 5 Schematically illustrated is a simplified representation of a computing system 500 that is configured to provide any to all of the computing functionality described herein. The computing system 500 may take the form of one or more personal computers, network-accessible server computers, tablet computers, home entertainment computers, gaming devices, mobile computing devices, mobile communication devices (e.g., smart phones), virtual / augmented / mixed reality computing devices, wearable computing devices, Internet of Things (IoT) devices, embedded computing devices, and / or other computing devices. For example, the computing system 500 may include any combination of logic subsystems, storage subsystems, and / or other subsystems of one or more of the search engine computing system 100, the client computer 102, the search result provider 110, the identity data repository 124, and the experience database 130.

[0074] The computing system 500 includes a logic subsystem 502 and a storage subsystem 504. The computing system 500 may optionally include an input / output subsystem 506 (e.g., including one or more input devices or sensors, and one or more output devices, such as a graphics display and / or audio speakers), a communication subsystem 508, and / or Figure 5 Other subsystems not shown.

[0075] The logic subsystem 502 includes one or more physical devices configured to execute instructions. For example, the logic subsystem may be configured to execute instructions as part of one or more applications, services, or other logical structures. The logic subsystem may include one or more hardware processors configured to execute software instructions. Additionally or alternatively, the logic subsystem may include one or more hardware or firmware devices configured to execute hardware or firmware instructions. The processor of the logic subsystem may be single-core or multi-core, and the instructions executed thereon may be configured for sequential, parallel, and / or distributed processing. The various components of the logic subsystem may optionally be distributed in two or more separate devices, which may be remotely located and / or configured for collaborative processing. Various aspects of the logic subsystem may be virtualized and executed by a remotely accessible networked computing device configured in a cloud computing configuration.

[0076] The storage subsystem 504 includes one or more physical devices configured to temporarily and / or permanently hold computer information, such as data and instructions that can be executed by the logic subsystem. When the storage subsystem includes two or more devices, these devices can be located in parallel and / or remotely. The storage subsystem 504 can include volatile, non-volatile, dynamic, static, read / write, read-only, random access, sequential access, location addressable, file addressable and / or content addressable devices. The storage subsystem 504 can include removable and / or built-in devices. When the logic subsystem executes instructions, the state of the storage subsystem 504 can be converted to, for example, hold different data.

[0077] Aspects of logic subsystem 502 and storage subsystem 504 may be integrated together into one or more hardware logic components. Such hardware logic components may include, for example, program-specific and application-specific integrated circuits (PASIC / ASIC), program-specific and application-specific standard products (PSSP / ASSP), systems on chips (SOC), and complex programmable logic devices (CPLD).

[0078] The logic subsystem and the storage subsystem can cooperate to instantiate one or more logic machines. As used herein, the term "machine" is used to collectively refer to hardware and any software, instructions, and / or other components that cooperate with such hardware to provide computer functions. In other words, "machine" is never an abstract concept and always has a tangible form. A machine can be instantiated by a single computing device, or a machine can include two or more subcomponents instantiated by two or more different computing devices. In some implementations, the machine includes a local component (e.g., a computer service) that cooperates with a remote component (e.g., a cloud computing service). The software and / or other instructions that give a particular machine its functionality can optionally be saved as an unexecuted module on a suitable storage device. Non-limiting examples of machines that can be instantiated by a computing system 500 according to the present disclosure include components of a web browser 106, a search result provider 110, a ranker 112, and a record pipeline 114.

[0079] Machines according to the present disclosure may be implemented using any suitable combination of existing and / or future machine learning (ML), artificial intelligence (AI), and / or natural language processing (NLP) techniques. Non-limiting examples of techniques that may be incorporated into implementations of one or more machines include support vector machines, multi-layer neural networks, convolutional neural networks (e.g., including spatial convolutional networks for processing images and / or videos, temporal convolutional neural networks for processing audio signals and / or natural language sentences, and / or any other suitable convolutional neural network that is configured to span one or more temporal and / or spatial dimensions), recurrent neural networks (e.g., long short-term memory networks), associative memories (e.g., lookup tables, hash tables, Bloom filters, neural Turing machines, and / or neural random access memories), word embedding models (e.g., GloVe or Word2Vec), unsupervised spatial and / or clustering methods (e.g., nearest neighbor algorithms, topological data analysis, and / or k-means clustering), graphical models (e.g., (hidden) Markov models, Markov random fields, (hidden) conditional random fields, and / or artificial intelligence knowledge bases), and / or natural language processing techniques (e.g., tokenization, stemming, constituency and / or dependency parsing, and / or intent identification, segmentation models, and / or hyper-segmentation models (e.g., hidden dynamic models)).

[0080] In some examples, the methods and processes described herein can be implemented using one or more differentiable functions, where the gradients of the differentiable functions can be calculated and / or estimated with respect to the inputs and / or outputs of the differentiable functions (e.g., relevant training data, and / or relevant objective functions). Such methods and processes can be determined, at least in part, by a set of trainable parameters. Thus, the trainable parameters of a particular method or process can be adjusted by any suitable training procedure to continuously improve the functionality of the method or process. For example, machine learning training techniques can be used to mine user approval / disapproval signals, e.g., to determine whether to add new query blacklist entries, site rules, and / or pattern rules for prohibiting query answers.

[0081] Non-limiting examples of training procedures for adjusting trainable parameters include supervised training (e.g., using gradient descent or any other suitable optimization method), zero-shot, few-shot, unsupervised learning methods (e.g., classification based on classes derived from unsupervised clustering methods), reinforcement learning (e.g., feedback-based deep Q-learning), and / or generative adversarial neural network training methods, belief propagation, RANSAC (random sample consensus), contextual bandit methods, maximum likelihood methods, and / or expectation maximization. In some examples, multiple methods, processes, and / or components of the systems described herein can be trained simultaneously with respect to an objective function that measures the performance of the collective function of multiple components (e.g., with respect to reinforced feedback and / or with respect to labeled training data). Training multiple methods, processes, and / or components simultaneously can improve such collective function. In some examples, one or more methods, processes, and / or components can be trained independently of other components (e.g., offline training on historical data).

[0082] The methods and processes disclosed herein can be configured to give users and / or any other person control over any privacy and / or potentially sensitive data. Whenever data is stored, accessed, and / or processed, it can be processed in accordance with privacy and / or security standards. When collecting user data, users or other stakeholders can specify how the data is used and / or stored. Whenever user data is collected for any purpose, the user who owns the data should be notified, and user data should only be collected if the user provides affirmative consent. If data is to be collected, it can and should be collected with the greatest respect for user privacy. If the data is to be accessed by anyone other than the user or used in any decision-making process, the user's consent may be obtained before the data is used and / or released. Users can opt in and / or opt out of data collection at any time. After the data is collected, the user can issue a command to delete the data, and / or restrict access to the data. All potentially sensitive data can be optionally encrypted and / or anonymized where feasible to further protect user privacy. Users can specify partial data, metadata, or statistics / results of processing data to be released to other parties, for example, for further processing. Private and / or confidential data may be kept completely confidential, e.g., only temporarily decrypted for processing, or only decrypted for processing on a user device and otherwise stored in encrypted form. The user may hold and control the encryption keys for the encrypted data. Alternatively or additionally, the user may designate a trusted third party to hold and control the encryption keys for the encrypted data, e.g., to provide the user with access to the data in accordance with an appropriate authentication protocol.

[0083] When the methods and processes described herein are incorporated into ML and / or AI components, the ML and / or AI components may make decisions based, at least in part, on the training of the components on training data. Thus, the ML and / or AI components may and should be trained on diverse, representative data sets that include sufficient relevant data for different users and / or user groups. In particular, the training data sets should contain diverse human individuals and groups so that as the ML and / or AI components are trained, performance in terms of user experience for the users and / or user groups is improved.

[0084] For example, a dialog system according to the present disclosure may be trained to interact with different groups of users, using language models trained to work well with those populations based on the language, dialect, accent, and / or any other features of the speaking style of the population.

[0085] ML and / or AI components can also be trained to make decisions that minimize potential bias against human individuals and / or groups. For example, when AI systems are used to assess any qualitative and / or quantitative information about human individuals or groups, they can be trained to be invariant to differences between individuals or groups that are not intended to be measured through qualitative and / or quantitative assessments, for example, so that any decisions are not accidentally influenced by differences between individuals and groups.

[0086] ML and / or AI components can and should be designed to provide context about how they operate whenever possible so that implementers of ML and / or AI systems can be held accountable for the decisions / evaluations made by the systems. For example, ML and / or AI systems should have replicable behavior, e.g., when they make pseudo-random decisions, a random seed should be used and recorded so that the decisions can be replicated later. As another example, data used to train and / or test machine learning and / or AI systems should be collated and maintained to facilitate future investigations of the machine learning and / or AI system's behavior in relation to the data. Furthermore, machine learning and / or AI systems can and should be monitored on an ongoing basis to identify potential bias, errors, and / or unexpected results.

[0087] When included, the input / output subsystem 506 can be used to present a visual representation of the data maintained by the storage subsystem 504. The visual representation can take the form of a graphical user interface (GUI). The input / output subsystem 506 can include one or more display devices using almost any type of technology. In some implementations, the input / output subsystem 506 can include one or more virtual reality, augmented reality, or mixed reality displays. The input / output subsystem 506 can be used to visually present content, such as a browser 106 and search results displayed in a page of the browser 106. The input / output subsystem 506 can include one or more microphones and / or speaker devices, which are configured to receive and / or output audio. In some examples, a microphone device can be used to receive voice audio input, which can be processed (e.g., using natural language processing and / or machine learning techniques) to receive user queries, determine user intent, etc. For example, voice audio input can be processed to control the browser 106. For example, voice audio input can be processed to identify a user's query to a search engine, for example, in addition to or in place of a user's text input in a search bar. In some examples, a speaker device may be used to output voice audio, such as providing information to a user, interacting with a user in a verbal conversation, etc. In some examples, the browser 106 may be configured to present content in the form of voice audio. For example, the browser 106 may present search results by outputting voice audio indicating the result entry for each result entry in the search results. For example, when the browser 106 presents search results including a query answer and a plurality of other result entries, the browser 106 may output voice audio that repeats the query answer, and output further voice audio that lists the title and / or summary of each result entry in the plurality of other result entries.

[0088] When included, the input / output subsystem may further include one or more input devices or interface with one or more input devices. The input device may include a sensor device or a user input device. Examples of user input devices include a keyboard, a mouse, a touch screen, or a game controller. In some embodiments, the input subsystem may include a selected natural user input (NUI) component or interface with a selected natural user input (NUI) component. Such components may be integrated or peripheral, and the conversion and / or processing of input actions may be processed on-board or off-board. Example NUI components may include microphones for voice and / or speech recognition; infrared, color, stereo and / or depth cameras for machine vision and / or gesture recognition; head trackers, eye trackers, accelerometers and / or gyroscopes for motion detection and / or intent recognition.

[0089] When included, the communication subsystem 508 can be configured to communicatively couple the computing system 500 with one or more other computing devices. The communication subsystem 508 can include wired and / or wireless communication devices compatible with one or more different communication protocols. The communication subsystem can be configured to communicate via a personal, local area network, and / or wide area network.

[0090] The language model can use vocabulary features to guide sampling / searching words to recognize speech. For example, the language model can be defined at least in part by the statistical distribution of words or other vocabulary features. For example, the language model can be defined by the statistical distribution of n-grams, defining the transition probabilities between candidate words based on vocabulary statistics. The language model can also be based on any other appropriate statistical features and / or the results of processing statistical features with one or more machine learning and / or statistical algorithms (e.g., confidence values ​​generated by such processing). In some examples, the statistical model can limit which words can be recognized for the audio signal, for example based on the assumption that the words in the audio signal are from a specific vocabulary.

[0091] Alternatively or additionally, the language model can be based on one or more neural networks previously trained to represent audio input and words in a shared latent space, for example, by one or more audio and / or word models (e.g., wav2letter and / or word2vec). Thus, finding a candidate word can include searching the shared latent space based on a vector encoded by the audio model to obtain the audio input in order to find a candidate word vector for decoding using the word model. For one or more candidate words, the shared latent space can be used to evaluate the confidence that the candidate word has a feature in the speech audio.

[0092] The language model can be used in conjunction with an acoustic model, which is configured to be based on acoustic features of the word (e.g., Mel frequency cepstral coefficients, formants, etc.) for candidate words and audio signals. Optionally, in some examples, the language model can be incorporated into the acoustic model (e.g., the evaluation and / or training of the language model can be based on the acoustic model). The acoustic model defines a mapping between an acoustic signal and a basic sound unit (such as a phoneme), for example, based on labeled speech audio. The acoustic model can be based on any suitable combination of existing or future machine learning (ML) and / or artificial intelligence (AI) models, such as: deep neural networks (e.g., long short-term memory, time convolutional neural networks, restricted Boltzmann machines, deep belief networks), hidden Markov models (HMMs), conditional random fields (CRFs) and / or Markov random fields, Gaussian mixture models, and / or other graphical models (e.g., deep Bayesian networks). The audio signal to be processed with the acoustic model can be preprocessed in any suitable manner, for example, encoded at any suitable sampling rate, Fourier transform, bandpass filter, etc. The acoustic model can be trained to identify the mapping between acoustic signals and sound units based on training on labeled audio data. For example, the acoustic model can be trained based on labeled audio data including speech audio and correction text to learn the mapping between the speech audio signal and the sound units represented by the correction text. Thus, the acoustic model can be continuously improved to improve its utility for correctly identifying speech audio.

[0093] In some examples, the language model can incorporate any suitable graphical model, such as a hidden Markov model (HMM) or a conditional random field (CRF), in addition to a statistical model, a neural network, and / or an acoustic model. The graphical model can utilize statistical features (e.g., transition probabilities) and / or confidence values ​​to determine the probability of recognizing a word, given the speech audio and / or other words recognized to date. Thus, the graphical model can utilize statistical features, previously trained machine learning models, and / or acoustic models to define transition probabilities between states represented in the graphical model.

[0094] In an example, a method for selectively providing different types of search engine results to different searchers includes: for each of a plurality of different unique identities, recording a browsing history for the unique identity, for each of a plurality of different rich snippet experiences configured to present selected snippet-specific information on a search results web page with other search results, associating the unique identity with the rich snippet experience in response to the browsing history for the unique identity satisfying a relevance criterion for the rich snippet experience, presenting the rich snippet experience on the search results web page in response to receiving a search query from the unique identity previously associated with the rich snippet experience, and presenting other search results without the rich snippet experience on the search results web page in response to receiving a search query from a different unique identity not previously associated with the rich snippet experience. In this example and / or other examples, the rich snippet experience can be presented with the other search results only if the search query received from the unique identity exceeds a relaxed trigger threshold. In this example and / or other examples, the method further includes presenting the rich snippet experience with the other search results if the search query received from the different unique identity exceeds a strict trigger threshold that is greater than the relaxed trigger threshold. In this example and / or other examples, if the search query received from the unique identity does not exceed the relaxation trigger threshold, the rich fragment experience may not be presented with other search results. In this example and / or other examples, if the search query received from the unique identity does not exceed the relaxation trigger threshold, the rich fragment experience may not be presented with other search results. In this example and / or other examples, recording the browsing history for the unique identity may include generating a keyword tag derived from one or more of the following: a previous search query, a previously selected search result, and a previously presented rich fragment experience. In this example and / or other examples, each rich fragment experience may be defined according to a different set of keyword tags, and wherein the relevance criteria are met based on the keyword tag set associated with the rich fragment experience and the keyword match associated with the unique identity. In this example and / or other examples, recording the browsing history for the unique identity may include: filtering out previously selected search results with a dwell time less than a threshold time so as not to be used to generate keyword tags for the unique identity. In this example and / or other examples, the method also includes: generating a confidence score for the rich fragment experience based on the search query, and wherein the rich fragment experience is positioned on the search result webpage based on the confidence score. In this example and / or other examples, other search results can be personalized for the unique identity based on the browsing history for the unique identity. In this example and / or other examples, the unique identity can be associated with a user identity across different computing devices. In this example and / or other examples, the unique identity can be associated with a specific computing device.

[0095] In one example, a method for selectively providing different types of search engine results includes: recording a browsing history for a unique identity, for each rich snippet experience in a plurality of different rich snippet experiences configured to present featured snippet-specific information on a search results web page together with other search results, in response to the browsing history of the unique identity that meets a rich snippet experience relevance standard, associating the unique identity with a rich snippet experience, in response to receiving a search query from the unique identity, presenting the rich snippet experience on the search results web page together with the other search results based on rich snippet experiences previously associated with the unique identity and rich snippet experiences that exceed a relaxed trigger, presenting other search results on the search results web page without the rich snippet experience based on the rich snippet experience being previously associated with the unique identity and not exceeding a relaxed trigger threshold, presenting the rich snippet experience on the search results web page together with the other search results based on rich snippet experiences that were not previously associated with the unique identity and rich snippet experiences that exceed a strict trigger threshold, and presenting the other search results on the search results web page without the rich snippet experience based on the rich snippet experience not being previously associated with the unique identity and the rich snippet experience not exceeding the strict trigger threshold. In this example and / or other examples, recording the browsing history for a unique identity may include generating a keyword tag derived from one or more of: a previous search query, a previously selected search result, and a previously presented rich fragment experience. In this example and / or other examples, each rich fragment experience may be defined according to a different set of keyword tags, and wherein the relevance criteria are met based on a match between the keyword tags associated with the unique identity and the keyword tag set associated with the rich fragment experience. In this example and / or other examples, recording the browsing history for a unique identity may include filtering out previously selected search results with a dwell time less than a threshold time so as not to be used to generate keyword tags for the unique identity. In this example and / or other examples, the method further includes: generating a confidence score for the rich fragment experience based on the search query, and wherein the rich fragment experience is positioned on the search result web page based on the confidence score.

[0096] In one example, a method for selectively providing different types of search engine results to different searchers includes: for each of a plurality of different unique identities, recording a browsing history for the unique identity, for each of a plurality of different rich snippet experiences configured to present selected snippet-specific information on a search results web page along with other search results, in response to the browsing history of the unique identity satisfying a relevance standard for the rich snippet experience, associating the unique identity with the rich snippet experience, in response to receiving a search query from a unique identity previously associated with the rich snippet experience, presenting the rich snippet experience on the search results web page if the search query received from the unique identity exceeds a relaxed trigger threshold, presenting other search results without the rich snippet experience on the search results web page if the search query received from the unique identity does not exceed the relaxed trigger threshold, in response to receiving a search query from a different unique identity not previously associated with the rich snippet experience, presenting the rich snippet experience along with the other search results if the search query received from the different unique identity exceeds a strict trigger threshold that is greater than the relaxed trigger threshold, and presenting the other search results without the rich snippet experience on the search results web page if the search query received from the different unique identity does not exceed the strict trigger threshold. In this example and / or other examples, recording the browsing history for a unique identity may include: generating a keyword tag derived from one or more of: a previous search query, a previously selected search result, and a previously presented rich snippet experience. In this example and / or other examples, each rich snippet experience may be defined according to a different set of keyword tags, and wherein the relevance criteria are met based on a match between the keyword tags associated with the unique identity and the set of keyword tags associated with the rich snippet experience. In this example and / or other examples, recording the browsing history for a unique identity may include: filtering out previously selected search results with a dwell time less than a threshold time so as not to be used to generate keyword tags for the unique identity.

[0097] It should be understood that the configurations and / or methods described herein are exemplary in nature, and these specific embodiments or examples should not be considered restrictive, as many variations are possible. The specific procedures or methods described herein may represent one or more of any number of processing strategies. Therefore, the various actions shown and / or described may be performed in the order shown and / or described, in other orders, in parallel or omitted. Likewise, the order of the above process may be changed.

[0098] The subject matter of the present disclosure includes all novel and nonobvious combinations and sub-combinations of the various processes, systems and configurations disclosed herein, and other features, functions, acts, and / or properties disclosed herein, as well as any and all equivalents thereof.

Claims

1. A method performed by a search engine, the search engine being used to generate a search engine results page (SERP) based on a query, the method include: receiving the query, wherein the query has a unique identity for a user or a device assigned to a user; identifying search results based on the query; generating a first keyword list based on the query and the search results identified based on the query; calculating a score for a rich snippet experience, wherein the rich snippet experience has a second keyword list assigned to it, and further wherein the score is calculated based on a keyword in the first keyword list matching a keyword in the second keyword list; When the unique identity is associated with the rich snippet experience and the score exceeds a relaxed trigger threshold, presenting the rich snippet experience with the search result on the SERP, wherein the relaxed trigger threshold is lower than a strict trigger threshold; as well as When the unique identity is not associated with the rich snippet experience and the score exceeds the strict trigger threshold, presenting the rich snippet experience along with the search result on the SERP, wherein the relaxed trigger threshold is lower than the strict trigger threshold.

2. The method of claim 1, wherein the rich snippet experience is information relevant to the query that is curated and presented in a visually distinctive manner relative to other non-curated search results.

3. The method according to claim 1, further comprising: include: calculating a second score for a second rich snippet experience, wherein the second rich snippet experience has a third keyword list assigned to it, and further wherein the second score is calculated based on keywords in the first keyword list matching keywords in the third keyword list; and Based on the second score, presenting the second rich snippet experience with the search result on the SERP.

4. The method according to claim 3, further comprising: include: Based on determining that the second score exceeds the strict trigger threshold, presenting the second rich snippet experience with the search result on the SERP. 5 . The method of claim 1 , further comprising recording a browsing history for the unique identity, wherein the rich snippet experience is assigned to the unique identity based on the browsing history for the unique identity. The method of claim 1 , wherein the rich snippet experience is positioned on the SERP based on the score. The method of claim 1 , wherein the search results are personalized for the unique identity based on a browsing history for the unique identity. The method of claim 1 , wherein the unique identity is associated with a user identity across different computing devices.

9. The method of claim 1, wherein the unique identity is associated with a computing device.

10. A computing system, include: processor; as well as a memory storing instructions that, when executed by the processor, cause the processor to perform actions comprising: receiving a query, wherein the query has a unique identity for a user or a device assigned to a user; identifying search results based on the query; generating a first keyword list based on the query and the search results identified based on the query; calculating a score for a rich snippet experience, wherein the rich snippet experience has a second keyword list assigned to it, and further wherein the score is calculated based on a keyword in the first keyword list matching a keyword in the second keyword list; When the unique identity is associated with the rich snippet experience and the score exceeds a relaxed trigger threshold, presenting the rich snippet experience with the search results on a search engine results page (SERP), wherein the relaxed trigger threshold is lower than a strict trigger threshold; and When the unique identity is not associated with the rich snippet experience and the score exceeds the strict trigger threshold, presenting the rich snippet experience along with the search result on the SERP, wherein the relaxed trigger threshold is lower than the strict trigger threshold.

11. The computing system of claim 10, wherein the rich snippet experience is information relevant to the query that is curated and presented in a visually distinctive manner relative to other non-curated search results.

12. The computing system of claim 10, wherein the action further include: calculating a second score for a second rich snippet experience, wherein the second rich snippet experience has a third keyword list assigned to it, and further wherein the second score is calculated based on keywords in the first keyword list matching keywords in the third keyword list; and Based on the second score, presenting the second rich snippet experience with the search result on the SERP.

13. The computing system of claim 12, wherein the action further include: Based on determining that the second score exceeds the strict trigger threshold, presenting the second rich snippet experience with the search result on the SERP.

14. The computing system of claim 10, wherein the action further include: A browsing history for the unique identity is recorded, wherein the rich snippet experience is assigned to the unique identity based on the browsing history for the unique identity.

15. The computing system of claim 10, wherein the rich snippet experience is positioned on the SERP based on the score.

16. The computing system of claim 10, wherein the search results are personalized for the unique identity based on a browsing history for the unique identity.

17. The computing system of claim 10, wherein the unique identity is associated with a user identity across different computing devices.

18. The computing system of claim 10, wherein the unique identity is associated with a computing device.

19. A non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform actions, the actions include: receiving a query, wherein the query has a unique identity for a user or a device assigned to a user; identifying search results based on the query; generating a first keyword list based on the query and the search results identified based on the query; calculating a score for a rich snippet experience, wherein the rich snippet experience has a second keyword list assigned to it, and further wherein the score is calculated based on a keyword in the first keyword list matching a keyword in the second keyword list; When the unique identity is associated with the rich snippet experience and the score exceeds a relaxed trigger threshold, presenting the rich snippet experience with the search results on a search engine results page (SERP), wherein the relaxed trigger threshold is lower than a strict trigger threshold; as well as When the unique identity is not associated with the rich snippet experience and the score exceeds the strict trigger threshold, presenting the rich snippet experience along with the search result on the SERP, wherein the relaxed trigger threshold is lower than the strict trigger threshold.

20. The non-transitory computer-readable medium of claim 19, wherein a second rich snippet experience and the search result are presented on the SERP simultaneously with the rich snippet experience.

Citation Information

Patent Citations

  • Context promotion of alternative search result

    CN102982120A

  • Systems and methods that enable search engines to present relevant snippets

    US20060161542A1

  • Personalizing a search results page based on search history

    US20070294615A1