KR20020070091A - User-profile driven mapping of speech onto URLs - Google Patents

Abstract

The Internet service provider, through the control of speech-enabled hyperlinks on the portal, allows the user to access the first web site with the first web address and the second web site with the second web address, respectively. 2 Allow parties to sell or rent specific speech entered into a speech-enabled browser. User access is provided to the first web site or the second web site depending on the geographical location of the user's client when the user activates the hyperlink through speech input.

Description

User-profile driven mapping of speech onto URLs}

The Internet is a vast resource of information that exists in web documents. Reportedly, 1.5 million web pages are added daily. Web documents typically include hyperlinks that allow movement from one web document to other web documents indicated by the network address of the hyperlinks. As a result, any document is connected to any other document via the appropriate root of hyperlinks, so its name is "Web". The importance of allowing user-friendly interaction with the Internet to selectively place specific web documents based on their information content cannot be overemphasized.

User-interaction is, for example, reliable and accurate search engines such as PlanetSearch of Philips Electronics, which allows users to query indexed inventory. ) Is facilitated. Planet Search's user interface provides each particular one of the query results with indications of the relative weight of each of the plurality of query terms contributed to the ranking of the particular result. See US Patent Serial No. 08 / 922,825 (Representative Document PHA21,935A), filed September 3, 97 and US Patent Serial No. 08 / 371,188, Representative Document PHA 21,935B, filed March 20, 98. , Both of which are incorporated herein by reference. In addition, U. S. Patent Serial No. 09 / 104,491 (Representative Document PHA 23.422), filed June 25, 98, is incorporated herein by reference, which includes information retrieval based on semantic context and user profiles. It is starting.

Another contribution to facilitating web browsing is based on speech recognition. US Patent Serial No. 09 / 174,167 (Representative Document PHN 16,571), filed Oct. 16, 98, incorporated herein by reference, discloses a client server architecture for applying speech recognition for interaction on the Internet. -sever architecture) is disclosed. The browser plug-in records the user's voice and pre-processes it to generate feature vectors. The feature vectors are sent to a speech recognition server via the internet.

More specifically, US Patent Serial No. 09 / 174,167 discloses a method and system for recognizing an input pattern stored at a user station and using a recognition unit of a server station. The server station and user station are connected via a network. The recognition unit is operative to recognize an input pattern using a model collection of at least one recognition model. The initial recognition registration step performs the transfer of model improvement data related to the user of the user station from the user station to the recognition unit and to associate the user identifier with the user of the user station. For a recognition session between a user station and a server station, the user station sends a user identifier associated with the user of the user station to the server station. The input pattern is also sent to a server station representing a temporally continuous input generated by the user. The recognition unit is used to recognize the input pattern by integrating at least one recognition model in the model collection that reflects model improvement data associated with the user. The server includes a plurality of different recognition models of the same type. In the recognition registration step, the recognition model is selected from a plurality of different recognition models of the same type, and is selected according to model improvement data related to the user. Further, in the registration step, the indication of the selected recognition model is stored in association with the user identifier. Further, in the step of recognizing the input pattern, the recognition model is retrieved in association with the user identifier passed to the server station, and the retrieved recognition model is integrated into the model collection.

By storing a plurality of recognition models of the same type, ie, a plurality of language models each targeting at least one other subject such as photography, gardening, automobile, etc., an appropriate recognition model can be selected for a particular user of the system. This allows a good perception. In this way, the user is not limited to one particular type of recognition model, such as a specific language model or vocabulary. On the other hand, system compatibility is achieved by reusing models for many users at the same time. For example, all users interested in photography can use the same language model to cover photography. Thus, the relevant good recognition results provided by the use of a user-oriented recognition model and this compatibility are achieved without storing a specific model for each user. Advantageously, the amount of training data required to be provided by the user can also be substantially less than in known systems. Instead of requiring a sufficient amount of data to fit an existing model or to train the model as a whole, according to the present invention, only a sufficient amount of data is needed to select the appropriate model from the available models. Shall be.

Now, with the speech technique described above, speech-activated hyperlinks can be implemented to enable speech-based Web browsing. Speech input from the user is handled so that it functions the same as clicking the hyperlink with a mouse. Speech input is supplied by a recognizer. The recognizer provides the recognition result and supplies the result to a database of stored words. The database has hyperlinks associated with stored words. For example, a particular word is associated with a particular hyperlink or set of synonyms, and semantically similar words or similar expressions are all associated with specific hyperlinks. If there is a match, the associated hyperlink is activated and the user accesses a new web site whose URL is identified in this hopper link.

For example, when registering as a service provider, a person can freely associate URLs with specific words or expressions (eg, "my picture", "my family" ...) on that person's homepage. . The service provider supplies some of the HTML code that the person places on his / her web (home) pages. Each time this page is downloaded by someone else, the code triggers a speech button to be displayed in the browser. The user presses the button and enters a word or sequence of words through speech. When the button owner returns a word that specifies the URL, the user connects to the corresponding new web page. Also, if the URL addresses a (cgi) script, the homepage can be created dynamically. The vocabulary of the recognizer also includes words that the service provider itself can associate with the link. If such (general) words such as "books" are entered as speech input, the system will, for example, provide the user with a category page linked to related topics as well as (targeted) advertising banners. . Alternatively, words or expressions may be sold to a customer. For example, when a user enters the word "books" as speech, he or she connects to a company's web site that sells books over the Internet.

The invention relates in particular to browsing on the Internet via speech input.

1 is a block diagram illustrating a business model of the present invention.

Above all, it is an object of the present invention to provide a user-friendly way of browsing the web. Another object is to use additional degrees of freedom to hyperlink web sites, for example, to increase user-friendly interaction with the Internet.

The present invention relates to a method for enabling a user to navigate on the Internet. User-interaction with hyperlinks is possible through the browser on the client's client. The first website or the second website is provided to the user according to the user's profile when the user activates the hyperlink. The first and second web sites have different URLs or web addresses. The user profile is for example based on the user's IP address or on the user's client's hardware and / or software profile. The IP address typically includes an indication of the geographical location of the user's client. This location is then used to selectively link to the first or second web site.

If the browser is speech-controllable or speech-enabled, the hyperlink is speech-enabled via specific speech input to the browser. In connection with a mouse click, speech input adds additional size or degrees of freedom to the selection of the web address to be associated with the hyperlink. The mapping of speech input on a web address to the same hyperlink representative on the user's display monitor may depend on the user's geographical location as well as the language used so recognized.

The present invention also provides an Internet service. The service, through the control of speech-enabled hyperlinks on a portal, allows a first party to reserve a particular speech input in a speech-enabled browser for user access of a first web site with a first web address. Make it possible. The service also allows, through the control of hyperlinks on the portal, the second party to reserve the same specific speech input in the browser for user access of the second web site with the second web address. User access is provided to the first web site or the second web site depending on the user's profile when the user activates the hyperlink through speech input. The profile may depend on the IP address of the user to which the geographical area of the user's client can be sent.

The invention also provides a new business model for Internet services based on leasing, licensing or sales agreements between web addresses and user profiles. The service makes it possible to provide user access to portals with hyperlinks. The first web address of the first party is assigned to the hyperlink. In addition, the second web address of the second party is assigned to the same hyperlink. User access to the first web address or the second web address depends on the user's profile when the user activates the hyperlink. Again, the profile depends, for example, on the geographic area of the user's client.

The inventors are now proposed to consider the user's profile, for example the geographical area of the user's client device. This geographical area of the user can be derived from his / her IP address. The database described above now has an entry for a geographic location or region, adding stored words, hyperlinks and relationships between them. The particular hyperlink that is activated is a function of one or more query words and the geographic location of the user's client. This keyword-location combination can now activate certain hyperlinks associated with, for example, geographic regions. When using the keyboard or speech to enter, for example, "restaurants", CDs "or" bicycle roads ", the user will be able to enter each local restaurant, music store or bike path path belonging to his or her geographic area. You get to go to the listed websites.

The present invention supports a new business model for selling or leasing hyperlinks, especially speech active hyperlinks, to companies or institutions for a geographic area. For example, regardless of his / her location, or through speech-enabled hyperlinks across the globe, instead of users searching for the same web site on the go, the user now has a hyperlink associated with his / her location. You can search for. This geographic choice is of interest to businesses and institutions with only local interest, such as theaters, restaurants, shops, and to Internet service providers. Linguistic differences or dialects may be treated as considerations for determining the geographic location of a user.

Thus, when users type or speak "flower shop", they can find their adjacent shop. The present invention can tailor the response of the recognition system to everyone's stored preferences and regions. As mentioned above, the user's location can be easily retrieved from his / her IP address. Commands for keyword position combinations are stored in the database. Localization may be indicated, for example, by a zip code. The user's category page can be dynamically established, including links to local businesses as well as links to global companies, for example, mail order shops or online service providers operating nationally or globally. Can be. Also desired advertising banners may be placed. Alternatively, if the word is recalled, other companies may be associated with other locations. For example, a company may sell or lease words for a set of postal codes. Even a mixed approach is possible. For some users, the word can lead to a (local) category page, while others can take the appropriate company's site.

The user's preferences can be used to determine what action to take when a word is recognized. For this purpose, it is essential that the user's plug-in has a unique identifier. Also, if a cookie can be a user or if user verification is accurate enough, this authentication can be used as another alternative. The user may register his / her preferences with the service provider, or the user profile may be set by tracking his / her most popular words and Web surfing behavior. The preferences and profile information may include gender, name and hobbies, nationality, and the like. See also, for example, US Pat. No. 09 / 104,491 (Representative Document PHA 23.422), supra. Similar to the location information, this information can be used to determine which new page the user should take when a word is recognized.

In the above, the IP address is referred to as an example of user profile information that can be used to dynamically adapt the selection of hyperlinks. In addition to or in addition to considering IP address information, the setup of the user's client (eg, browser type and version, operating system, connection type) may also be used for dynamic suitability. Server-client interaction allows the server to extract this kind of setup information. Alternatively, the setup information is stored in a server obtained through a user registering with a service provider, for example. In this context, for example, the U.S. Patent, filed Jan. 17, 97, by Mehran Moshfeghi, Jun Wang, Stephen Wong and Yuan-Pin Yu, entitled "A method for personalizing hospital intranet web sites" See serial number 08 / 785,459 (agent document number PHA 23.217), which is incorporated herein by reference. This document relies on stored information indicative of a user's needs and preferences, including the user's needs and preferences assumed from stored information about the user's job, occupation or hospital's purpose, and usage profiles that have slowed motion. , A layer that dynamically creates web pages and other data objects using scripts such as graphics, audio and video files, a user's specific level of access to confidential patient information, and the computer and physical environments of the consumer, It's about the system. Obviously, the content is generated depending on the display resolution and the lowest bandwidth link between the server and the browser to limit not only the server load but also the latency for downloads. Now, if the user enters the word "software" in the form of a document or speech input, and the server system determines that the client is for example MacIntosh, the selection of software related sites is retrieved relative to the MacIntosh client. If the server determines that the client is an IBM compatible PC, another selection of software related to the documents is retrieved.

Hereinafter, the present invention will be described as an example with reference to the accompanying drawings.

The present invention allows a party to direct a user to a particular web site on the Internet. The portal is provided with a plurality of speech-controllable hyperlinks. Interaction with hyperlinks is possible through speech-enabled browsers. In a commercial transaction, a party is assigned a right to at least temporarily have at least one particular conventional language expression that invokes one particular speech control hyperlink to a particular web site. The web address of a particular web site retrieved via a hyperlink depends on the Internet Protocol (IP) address of the user's client. The geographic location of the client's client can be extracted from the user's IP address. An IP address typically contains four numbers (A, B, C, D). Typically, the last one (D) is the most specific / geographic and the least is the most general. Companies such as "double-click" collecting IP addresses store A; A.B.; A.B.C or even A.B.C.D relationships in their database. Some computers have fixed IP addresses and they do not change. Some clients (mostly individual ones) are connected to the Internet on request. They call their service providers and are assigned an IP address. This access point dynamically assigns IP addresses to their clients. They will have some prefix A.B or A.B.C and the last numbers C.D or D are assigned dynamically, respectively. By knowing where the access point with the prefix A.B or A.B.C will be located, we can assume the location where the client is located. These prefixes are stable and can be stored in a database. The geographical area location of the client can be determined dynamically and can be real time. Thus, a representation may be sold, leased and licensed by a geographic area to a commercial entity or institution with invented interests within that area. For example, the phrase may be used as a portal or gateway's speech-enabled hyperlinks, ie, a web site providing a gateway to the rest of the Internet, to direct users to web sites related to their geographic area. have.

1 is a block diagram illustrating the concept of a business model of the present invention. The first party 102 is interested in commercial applications local to the first geographic area and informs the public through his website. The second party 104 likewise concerns local but similar commercial utilization in a second geographic region different from the first region. The Internet service provider 106 controls the gateway, portal with various hyperlinks 108, of general interest to a broad audience. The portal is graphically represented on the visitor's client's display in a unique way, regardless of the geographic location of the client. The parties 102, 104 are interested in fascinating attention via this portal. The service provider 106 commercially utilizes his portal's screen real estate and does this through attention brokerage on a regional basis. Both parties 102 and 104 allow them to have a web address 11 of their respective web sites as associated with a particular of the hyperlinks 108 on the portal and associated with their respective geographic areas 114. Register as a lease unit (110). Now, a first client 116 residing in the first geographic area accesses the portal via the Internet 118, and a second client 120 residing in the second geographic area accesses the same portal via the Internet 118. Access. On their respective display monitors 117, 119, the users see the same graphical representation 121 of the portal. Both client devices activate the same graphical representation 123 of hyperlinks for parties 102 and 104 to register their lease. Based on the IP addresses of the clients 116, 120, as described above, this may be determined within the geographic area in which they reside. Based on the determined area, the service provider 106 automatically provides the client 116 with access to the party's 102 website, and accesses the party's 104 website only when the hyperlink is clicked. Is automatically provided to the client 120.

Now assume that clients 116 and 120 have speech-controllable browsers 122 and 124, respectively. The portal's hyperlinks 108 are speech-activated through the speech recognition system, for example through the infrastructure of the system as disclosed in U.S. Patent Serial No. 09 / 174,167 (representative document PHN 16,571). . Speech input of the same or different words can now be used to access other web sites depending on the geographical location of the requesting client. The service provider 106 leases the speech input and registers the words 126 assigned to the parties 102, 104 associated with their web address. The service provider 106 can now rent to the interested parties 102, 104 various combinations, that is, portal property (distributed) in the form of hyperlinks, other geographic areas and other or the same words. . Thus, the possibility of attention brokerage is improved via the portal 121.

Reference is made to PCT International Application International Publication No. WO 99/12104, entitled "Method and arrangement for finding information". This publication discloses a communication system for mobile stations that can retrieve web pages under the control of the geographical location of the mobile user. This location information is supplied by the mobile telephone system. Geographic information is sent from the telephone system to the mobile station. The user of the mobile station establishes a connection with the Internet and sends a request for documents. This request includes location information. The content of the web pages returned for that request is related to the geographical location of the mobile user as determined by the mobile phone system. Execution of this known method involves the user executing a Domain Name System (DNS) query. DNS is a service that translates domain names (eg www.philips.com) into IP addresses. The user writes a predetermined text in the browser window. This phrase is sent to an Access Node that provides access to the Internet. The access node requests and receives a geographical location as a mobile user from the mobile telephone system. The access node then combines the DNS query with the geographic information, re-searching the user by mapping the combination on the IP address and sending the IP address according to the translation table. The user can then launch a browser for the query based on the received IP address.

Claims (8)

In a method for enabling a user to navigate on the Internet, The method, Enabling user-interaction with the hyperlink 123 via the browser 122, 124 at the user's client 120, 116, The browser is speech-controllable, The hyperlink is speech-enabled via a specific speech input to the browser, When the user activates the hyperlink, providing a first website or a second website to a user that depends on the user's profile. The method of claim 1, The user profile is dependent on the user's IP address. The method of claim 1, And the user profile comprises an indication of the geographic location of the client of the user. The method of claim 1, The user profile depends on at least a hardware profile or a software profile of the client of the user. The method of claim 1, The user profile relies on a history of user-interaction with the internet via the speech-enabled browser. In the Internet service providing method, The method, Controlling the speech-enabled hyperlink on the portal 121 to allow the first party 102 to reserve specific speech input in a speech-enabled browser for user access to the first website with the first web address; , Allowing a second party 104 to reserve specific speech input in the browser for user access of a second web site having a second web address through control of a hyperlink on the portal; When the user activates the hyperlink through speech input, providing the user access to the first web site or the second web site depending on the user's profile. . In the Internet service providing method, The method, Providing user access to the portal (121) with speech-enabled hyperlinks (123) by speech input via speech-controllable browsers (122, 124). A user profile controllable and speech-activatable hyperlink on a web site as used in the method according to claim 1.