Patent Application
20100257019
Existing systems such as search engines provide information based on descriptions received from a user. The search engines infer intent based on the received descriptions, and provide the information based on the inferred intent. For example, if the user types “weather redmond wa” as a search query, the search engines infer that the user is interested in a forecast for the city of Redmond, Wash. The search engines might obtain and provide a five-day forecast within the search results along with the other links.
Existing systems, however, fail to consistently and accurately infer the intent of search queries at least because of the numerous search queries that may correspond to the same intent. For example, the search queries “Redmond forecast”, “is it going to rain tomorrow in Redmond”, and the like may all correspond to the same intent to obtain a forecast for Redmond. While query logs provide some insight into intent, the hundreds or thousands of search queries in the query logs in existing systems have to be manually labeled to extract meaningful data from them.
Embodiments of the invention identify descriptions for association with objects. A plurality of the objects is defined. Each of the objects is obfuscated. One or more of the obfuscated objects are provided to a plurality of users. Each of the users receives one of the obfuscated objects. The users each create the descriptions based on the provided objects. Each of the users reviews the descriptions from the other users. Each of the users makes a determination as to whether the users were provided the same objects. Associations between the descriptions and the provided objects are adjusted based on the determinations.
This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.
Corresponding reference characters indicate corresponding parts throughout the drawings.
Referring to the figures, embodiments of the disclosure enable, at least, the collection of object-to-description mappings 108. For example, a set of intent-to-query mappings may be collected in a search engine embodiment. Such a mapping enables the intent behind a particular search query to be inferred. As shown in
The data gathered by aspects of the disclosure may be used to learn a grammar or a set of linguistic patterns of how people express intentions in search queries 312. For example, while “what is the weather like in Seattle” and “weather forecast in Seattle” are search queries 312 for finding out about the weather in Seattle, these same linguistic patterns may be used to detect the intention of seeking information about the weather of any other cities. Knowing the intent of a search query allows for more intelligent and targeted ways of retrieving relevant search results (e.g., with fewer query reformulations) thereby enhancing the user experience by providing a complete set of results limited to the intent of the search query.
While embodiments of the disclosure describe the descriptions and objects with reference to search queries 312 and search intent, aspects of the disclosure are not limited to a search embodiment. Rather, other examples include: (1) human-generated written or voice descriptions of driving directions from one address to another to inform an automated system that provided directions, and (2) human-generated written or voice descriptions of images to inform an image-search system that used such descriptions as input.
Referring next to
The determinations are received at 210. If all the determinations are correct at 212, associations between the created descriptions and the provided object are defined or adjusted at 214. In a two-user example, if the same object was provided to both users 104 and both users 104 guessed this correctly, the created descriptions are associated with the object. If the created descriptions are already associated with the object, the ranking or weighting of the created descriptions is adjusted to indicate a greater association between the descriptions and the object. Conversely, if different objects were provided to both users 104 and both users 104 guessed this correctly, any association between the first object and the description for the second object, or between the second object and the description for the first object, is adjusted to indicate less of an association.
If not all the determinations are correct at 212, no adjustment is made at 216.
In some embodiments, the determinations from the users 104 include a value representing a quantity of the objects believed by each user 104 to have been provided. For example, in a two-player game, each user 104 makes a guess as to whether one or two objects (e.g., the same or different objects, respectively) have been provided. In such embodiments, the operations performed at 212 include determining the actual quantity of the provided objects and comparing the determined quantity to the value provided by each user 104.
Alternatively, the determinations from the users 104 include a different form of indication as to whether each user 104 has received the same or different objects. For example, each user 104 may type in the text “same” or “different” or select such a button or checkbox in a user interface. In such embodiments, the operations performed at 212 include comparing the text or selection from each user 104 with the correct determination.
In some embodiments, the operations performed at 214 include ranking the associated descriptions for the provided object. The ranking occurs based on various factors including, for example, a reputation of the user 104 creating the descriptions. In such an example, the descriptions from users 104 with higher reputations are ranked higher or weighted more than descriptions from users 104 with lower reputations.
Referring next to
The memory area 308 further stores a correlation between each of the search intentions 310 and one or more questions 314 and/or one or more instant answers 316. For example, if the search intention 310 is to identify the weather in Redmond, Wash., the correlated question 314 may be “What is the weather in Redmond, Wash.?” For the same search intention 310, the instant answer 316 may be “Redmond has light rain and 40 degrees.” In general, the instant answer 316 represents a concise result for some informational desire, and may include images, text, or other data.
In general, the memory area 308 is associated with the computing device 102. For example, in
The processor 306 is programmed to execute computer-executable instructions for implementing aspects of the disclosure. As an example, the processor 306 is programmed to execute instructions such as those illustrated in the figures (e.g.,
The memory area 308 further stores one or more computer-executable components. The components include an obfuscation component 318, an interface component 320, a search component 322, and a correlation component 324. The obfuscation component 318 generates a plurality of the questions 314 related to one or more of the search intentions 310.
In some embodiments, the obfuscation component 318 operates as follows. A set of predefined topics is selected (e.g., sports, medical procedures, companies, movies, celebrities, drug-condition interactions, products, etc.). For each topic, a set of question templates is created. Each question template is associated with a topic identifier and a question identifier. Questions 314 with the same topic identifier are similarly parameterized. For example, questions 314 about a particular drug (e.g., “what are the effects of Drug A?” and “what is the cost of Drug A?”) and questions 314 about the appropriateness of a drug for a particular condition (e.g., “can Drug A be taken during pregnancy?” and “can Drug A be taken by people with a heart condition?”) have two distinct topic identifiers. Questions 314 with the same topic and question identifiers are paraphrases of each other (e.g., “what is the cost of Drug A?” and “how much does Drug A cost?”) and are considered to be representations of the same search intention 310.
To construct questions 314 for the same search intention 310, questions 314 with the same topic and question identifier are randomly sampled. To construct questions 314 that are different but easily distinguishable, questions 314 may be sampled where the subjects are vastly different (e.g., drug versus movie, or female celebrity versus company) as shown by their different topic identifiers. Alternatively, questions 314 with the same topic and question identifiers, but with different entities substituted, may also be sampled.
To generate questions 314 that are different but more difficult to discriminate, questions 314 may be selected where the subjects are the same, but that the information inquired about the subjects is different. These are questions 314 with the same topic identifier but different question identifiers, and the same entity substituted. For example, the following are two question templates: “How much do the tickets cost for <game>?” and “Who won the <game>?”. For these templates, the tag <game>may be substituted with a specific game name. In this example, the players 304 are each given an intention to find some information about the specific game name. However, the particular kind of information sought (e.g., ticket price versus game result) differs. For the players 304 to tell that the questions 314 given to them are different, the players 304 judge from the search results that the questions 314 are about different aspects of the same subject.
In some embodiments, a mix of easy and difficult pairs of search intentions 310 is served in the game. The mix may be determined dynamically by observing the quantity of mistakes players 304 have made so far, and selecting the pairs of questions 314 accordingly to maintain a high level of player enjoyment.
The interface component 320 provides one or more of the questions 314 to the users 104. Each of the users 104 receives one of the questions 314. Each of the users 104 composes a search query 312 corresponding to the received question 314.
The search component 322 receives the search query 312 from each of the users 104. In some embodiments, the search component 322 performs a search on data using the received search queries 312 to generate search results. The search component 322 then provides the search results to the users 104. In other embodiments, another component (not shown) not associated with the computing device 102 performs the search and provides the search results to the users 104. In any embodiment, the users 104 are able to view the search results produced from their own search query 312 as well as the search results from the search queries 312 produced by the other user(s).
Each of the users 104 analyzes the search results and makes a determination. The determination indicates whether the user 104 believes that each of the users 104 has been provided with the questions 314 corresponding to the same search intention 310. The interface component 320 receives the determination from each of the users 104. Based on the determinations received by the interface component 320, the correlation component 324 adjusts an association between the search queries 312 and the search intentions 310. In some embodiments, the correlation component 324 defines an association between the search queries 312 and the search intention 310 if the determinations received from the users 104 are correct and if the provided questions 314 correspond to the same search intention 310. For example, the correlation component 324 compares the determinations to the known quantity of the search intentions 310 for which questions 314 were provided by the interface component 320. In embodiments, the correlation component 324 ranks or weights the search queries 312 such as described above with reference to
Referring next to
Each of the players 304 receives one of the questions 314. Each of the players 304 creates a search query 312 at 406 corresponding to the provided question 314. Based on the search queries 312, search results are obtained and provided to the players 304. For example, each of the players 304 receives the search results corresponding to their search query 312, as well as the search results corresponding to the search queries 312 from the other player(s). The players 304 review the search results at 408 to determine whether the players 304 received the same question 314.
In the two-player example, if both players 304 are correct in their determinations at 410, both players 304 are rewarded at 412. For example, if both players 304 correctly determine that the same question 314 was provided to both players 304, or that different questions 314 were provided to each player 304, both players 304 are rewarded. The reward may include any form of congratulations, accolades, or even compensation or credit. If, in the two-player example, either of the players 304 is incorrect in their determinations at 410, neither player 304 is rewarded at 414.
The exemplary operations illustrated in
In the two-player example, an example user interface for the game is next described with reference to
Referring next to
Referring next to
By way of example and not limitation, computer readable media comprise computer storage media and communication media. Computer storage media store information such as computer readable instructions, data structures, program modules or other data. Communication media typically embody computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism and include any information delivery media. Combinations of any of the above are also included within the scope of computer readable media.
Although described in connection with an exemplary computing system environment, embodiments of the invention are operational with numerous other general purpose or special purpose computing system environments or configurations. Examples of well known computing systems, environments, and/or configurations that may be suitable for use with aspects of the invention include, but are not limited to, mobile computing devices, personal computers, server computers, hand-held or laptop devices, multiprocessor systems, gaming consoles (including handheld gaming consoles), portable music players, a personal digital assistant, an information appliance, a personal communicator, microprocessor-based systems, set top boxes, programmable consumer electronics, mobile telephones, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
Embodiments of the invention may be described in the general context of computer-executable instructions, such as program modules, executed by one or more computers or other devices. The computer-executable instructions may be organized into one or more computer-executable components or modules. Generally, program modules include, but are not limited to, routines, programs, objects, components, and data structures that perform particular tasks or implement particular abstract data types. Aspects of the invention may be implemented with any number and organization of such components or modules. For example, aspects of the invention are not limited to the specific computer-executable instructions or the specific components or modules illustrated in the figures and described herein. Other embodiments of the invention may include different computer-executable instructions or components having more or less functionality than illustrated and described herein.
Aspects of the invention transform a general-purpose computer into a special-purpose computing device when configured to execute the instructions described herein.
The embodiments illustrated and described herein as well as embodiments not specifically described herein but within the scope of aspects of the invention constitute exemplary means for identifying a plurality of the search queries 312 that correspond to the same search intention 310, and exemplary means for obfuscating the search intentions 310.
The order of execution or performance of the operations in embodiments of the invention illustrated and described herein is not essential, unless otherwise specified. That is, the operations may be performed in any order, unless otherwise specified, and embodiments of the invention may include additional or fewer operations than those disclosed herein. For example, it is contemplated that executing or performing a particular operation before, contemporaneously with, or after another operation is within the scope of aspects of the invention.
When introducing elements of aspects of the invention or the embodiments thereof, the articles “a,” “an,” “the,” and “said” are intended to mean that there are one or more of the elements. The terms “comprising,” “including,” and “having” are intended to be inclusive and mean that there may be additional elements other than the listed elements.
Having described aspects of the invention in detail, it will be apparent that modifications and variations are possible without departing from the scope of aspects of the invention as defined in the appended claims. As various changes could be made in the above constructions, products, and methods without departing from the scope of aspects of the invention, it is intended that all matter contained in the above description and shown in the accompanying drawings shall be interpreted as illustrative and not in a limiting sense.
