WO2003073376A1

WO2003073376A1 - User identification methods and systems

Info

Publication number: WO2003073376A1
Application number: PCT/US2002/016674
Authority: WO
Inventors: Nathaniel Thurston; Devin Hosea
Original assignee: Predictive Media Corporation
Priority date: 2002-02-25
Filing date: 2002-05-29
Publication date: 2003-09-04
Also published as: AU2002322021A1

Abstract

A system and method, for use with a content delivery system (102) including a profiling process (112) operable to generate user profiles based on user actions, for determining whether a current user of the content delivery system is a previous user of the content delivery system. In one embodiment, the system receives a hitorical profile that represents a pattern of past user actions by at least one previous user. While being operated by the current user, the system compares groupings of most-recent user actions of the current user to the historical profile, and ones of the series of most-recent profile and the historical profile, and determines, on the basis of the calculated degrees of similarity, whether the current user is a previous user of the content delivery system. Using this method, the system may determine which of several viewers in a household is using the remote control (108) and/or viewing the programming, based on remote control interaction, viewing patterns and times of use.

Description

USER IDENTIFICATION METHODS AND SYSTEMS

PRIORITY CLAIM

The present application claims priority to co-pending Provisional Patent Application Serial No. 60/359,872 entitled User Identification Methods and Systems, filed on February 25, 2002, and having a common inventive entity. INCORPORATION BY REFERENCE

The present application for United States Patent incorporates by reference the following commonly-owned patent applications, as if set forth in their entirety herein, for all purposes:

WO 0120481A2 {Predictive Network PCT application};

U.S. Patent Application No. 60/338,398 filed December 7, 2001;

U.S. Patent Application entitled: "Television Program Navigation Guide" filed December 5, 2001;

U.S. Patent Application entitled: "Method and System for Selective Initial Television Channel Display" filed October 22, 2001;

U.S. Patent Application No. 09/969,911 filed October 3, 2001 ;

U.S. Patent Application entitled: "Method and System for Parsing Purchase Information from Web Pages filed August 29, 2001 ;

U.S. Patent Application No. 09/928,493 filed August 13, 2001; U.S. Patent Application No. 09/877,974 filed June 7, 2001 ;

U.S. Patent Application No. 09/558,755 filed April 21, 2001;

U.S. Patent Application No. 60/282,028 filed April 6, 2001 ;

U.S. Patent Application No. 09/798,337 filed March 2, 2001;

U.S. Patent Application No. 09/777,807 filed February 5, 2001 ; U.S. Patent Application No. 09/767,693 filed January 23, 2001 ; and

U.S. Patent Application No. 09/766,377 filed January 19, 2001. BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to methods and systems for identifying or distinguishing from among multiple users of a content distribution sys- tern, such as a television system equipped with an electronic or interactive programming guide (EPG/IPG).

2. Background

Electronic Program Guides (EPGs) are specialized graphical user interfaces (GUIs) that allow users to navigate through a multiple of pro- grams available for viewing or reproduction. Many EPGs present one or more layered displays — often available on a dedicated television channel — to give a user an inventory of programs available for viewing or reproduction, as well as schedules and other details (e.g., channels, program length, program synopsis, etc.) about the programs. Pointing devices ~ which may be incorporated into the remote control unit associated with the television monitor « are often used to control and navigate within the displays, as well as to activate links to the programs of interest. EPGs are especially useful for television viewers who receive a great deal of television stations, as is increasingly the case with cable and satellite television ser- vices.

The structure and operation of conventional EPG and IPG systems, as well as associated television systems, Internet systems, and subscriber devices, including set-top boxes (STBs), monitors, pointing devices, TV remote controls, and other control devices, are well-known in the art. Ex- amples are set forth in the U.S. and foreign patent documents listed above, the teachings of which are incorporated herein by reference as if set forth herein in their entirety.

FIG. 1 illustrates a conventional content delivery system 100 including a server 102 and client platform 104. Within the client platform 104 is a television monitor 106, remote control 108, and EPG or IPG system 110. Within the EPG system 110 there can be a profile engine 112, a recommendation engine 114 and a User Interface (UI) 116.

In a conventional EPG or IPG 110 like that shown in FIG. 1, the recommendation engine 114 rates each television show or other content available for viewing, using known methods described in the U.S. and foreign patent documents incorporated herein by reference. In particular, the recommendation engine 114 may use profile information made available by profile engine 112 to generate the ratings or recommendations. One or more user interfaces 116 make use of these ratings to assist the viewer in finding desired programming, often by generating ordered, ranked lists of shows and giving preferred placement within such lists to shows with higher ratings. The lists are then displayed to viewers using known user methods and devices in the UI layer 116. Many viewers rely on the ordered, ranked lists generated by such EPGs to select programs to be viewed. By way of example, using some prior art systems, a viewer might select an option such as "show highest ranked Sports program" for his or her next one-hour segment of viewing. EPGs thus have the potential to influence viewers' programming selections. Moreover, the ratings (and thus the rankings) assigned to shows by an EPG system can influence viewer- ship for such shows. Various configurations of EPGs and associated methods, systems and devices are disclosed in the following U.S. and foreign patent documents, among others, the disclosures of which are incorporated herein by reference as if set forth in their entirety here:

U.S. Pat. 6177931 Alexander et al.

WO 0049801A1 Yuen et al.

WO 0033224A1 Yuen

U.S. 6,163,316 Killian U.S. Pat. 6,005,597 Barrett et al.

More recently, EPG providers have begun to investigate the use of recommendation systems to enhance EPG/user interaction qualities. Such systems may, for example, provide content ratings and/or recommendations based on a user's viewing history, or his "profile" (a descriptive data- set that may be generated from the user's viewing history). Such features are helpful in enabling a user to quickly navigate to programming of interest. Within a household, however, there may be several users of the same content distribution system, each with markedly different viewing patterns. For example, mom, dad, son and daughter may all share use of the living room home entertainment system, but each may have markedly different interests, viewing patterns and preferred content. Even the genre preferred by each user may be different, ranging from sports to comedy to drama to horror. What is therefore desirable, but non-existent, are methods, devices and systems that would enable a content distribution system to identify or distinguish between different users of the content distribution system.

It is also desirable to provide such methods, devices and systems op- erable in real-time within or in association with an EPG environment.

It is also desirable to provide such methods, devices and systems suitable for implementation in a limited-resource platform such as a set-top box (STB).

SUMMARY OF THE INVENTION

The present invention provides methods, devices and systems that enable a content distribution system to identify or distinguish between different users of the content distribution system, in real-time, and thus provide appropriate ratings or recommendations for each user. The invention can be used in connection with a wide range of content delivery systems, including television systems and other media delivery systems using the Internet.

The user ID engine can quickly differentiate from among multiple family members who may share use of a home entertainment system in a household. By enabling rapid differentiation among users, the invention enables a content navigation/recommendation system incorporating the invention to instantly provide appropriate, personalized recommendations to the various users of the system. Sports fans will receive instantaneous recommendations for sports programming; aficionados of comedy will immediately see top comedy picks; and reality-show devotees will be quickly directed to their favorite content. These functions can be provided in real-time, and the invention's simplicity enables implementation on time, and the invention's simplicity enables implementation on compact platforms such as STBs.

One embodiment of the invention ~ suitable for deployment in a content delivery system including a profiling process operable to generate user profiles based on user actions ~ is a method of determining whether a current user of the content delivery system is a previous user of the content delivery system, including the following:

(1) during operation of the content delivery system by a current user, receiving from the profiling process at least one historical profile represen- tative of a pattern of user actions by at least one respective previous user;

(2) during operation of the content delivery system by the current user, receiving from the profiling process a series of most-recent profiles developed over groupings of most-recent user actions of the current user;

(3) iteratively comparing, to the at least first historical profile, ones of the series of most-recent profiles;

(4) for each comparison, calculating a degree of similarity value representative of a degree of similarity between a most-recent profile and the historical profile; and

(5) determining, on the basis of the calculated degrees of similarity, whether the current user is a previous user of the content delivery system.

This "biometrics" functionality of the invention thus enables the system to determine which of several viewers in a household is using the remote control and/or viewing the programming, based on remote control interaction, viewing patterns and times of use. BRIEF DESCRIPTION OF THE DRAWING FIGURES

Features and advantages of the present invention will become apparent to those skilled in the art from the description below, with reference to the following drawing figures, in which: FIG. 1 is a block diagram of a prior art content distribution system.

FIG. 2 is a block diagram showing a content distribution system incorporating a user ID engine or process in accordance with the invention. FIG. 3 is a flowchart illustrating method steps of one practice of the invention. FIG. 4 is a block diagram illustrating the processing of user action clickstream and profile information in accordance with one practice of the invention.

FIG. 5 is a diagram illustrating how the present invention handles a change of mode point; i.e., the point at which a current profile is converted into a historical profile.

DESCRIPTION OF ILLUSTRATED EMBODIMENTS

In the system shown in FIG. 1, ratings and recommendations are based on user profiles, such as those generated by profile engine 112. Since multiple users may share a living room home entertainment system or other content delivery system, the present invention provides methods and systems to distinguish, in real-time, from among multiple users of the system, and thereby enable the system to quickly offer appropriate recommendations to a user. This capability is based upon real-time analysis of user actions, such as button pushes on a TV remote control (channel up/down, volume up/down, and the like).

As shown in the block diagram of FIG. 2, one embodiment of the invention is a user ID engine or process 202 that communicates with profile engine 112 (which in turn may receive inputs from and/or transmit outputs to Remote Control 108, TV Monitor 106 and/or other components of EPG System 110 such as Recommendation Engine 118 and User Interface 116). The design and operation of Recommendation Engine 118 and User Interface 116 may be conventional in nature or may use the techniques described in the other applications and patents incorporated by reference herein. It will also be appreciated that although Profile Engine 112, Recommendation Engine 118, and User Interface 116 are shown in the diagram as part of the EPG System 110, those elements can also be separate from EPG System 110. The User ID engine can operate in conjunction with substantially any viewer profiling method disclosed in the prior art. See, for example, the profiling methods disclosed in the Yuen WIPO publications noted above, which are incorporated by reference herein in their entirety, as if set forth completely herein. Such methods typically generate viewer profiles by processing one or more streams of user events, such as remote control button pushes, channel up/down, volume up/down, and the like.

The operation of the user ID engine, in turn, is based on the princi- pie of comparing two profiles, generating a degree of similarity value, i.e., a metric of the probability that the two profiles were produced by the same behavior, and determining whether they were or were not, based on a threshold similarity value.

Referring now to FIG. 3, there is shown a flow chart illustrating the method of one practice of the invention. In the illustrated method, assuming a given profiling method, the following steps are executed:

First, the profiling system processes events (remote control button pushes and the like) to generate one or more profiles (step 302);

Next, the user ID engine compares profiles and generates degree-of- similarity metrics (step 304);

At all times, the user ID system maintains (step 306) a collection of:

(1) Historical profiles {H_i} and

(2) Several profiles representing current behavior, C_cur, with no time-weighting, and {C_i}, with an exponentially-decaying time- weighting. (For example, the time-weighting could have a half-life 2¹ for each i with 2' less than the length of the current profile;

At each iteration, the system (step 308) updates the current profiles and compares the behavior of each C_i with C_cur - C_i.

If the behavior is not consistent, then the system (at step 310): (1) Creates a new historical profile, H = C_cur - C_i (representing the portion of the current profile different from the current behavior - - his- torical profiles are merged if they represent similar behavior, so as to maintain a constant memory requirement); and

(2) Computes the degree of similarity between the current profile and the historical profile, using the aggregate profile: C_cur + sum i (H_i P(H_i given C_cur)) for computations reflecting the individual.

FIGS. 4 and 5 illustrate other aspects of the operation of this embodiment of the user ID engine. In particular, FIG. 4 shows the generation of current profiles 1, 2, ..., N by the profile engine, based on a received clickstream or other user actions, as well as the expiration of these profiles (due to their exponentially-decaying time-weighting), in X, X*2, . . ., and X*2^N, respectively. Also shown in FIG. 4 is the "current profile total", which has no expiration. It is this profile that is compared with the incremental profiles, with a frequency selected to be Y*2^N. (X and Y are tuning coefficients that can be selected by the implementer for effective function. In addition, other comparison frequencies or methods can be used, and are within the scope of the present invention.)

FIG. 5 shows that upon change of mode (step 500), as detected by the above-described comparisons and degree-of-similarity thresholding, a number of current profiles (denoted collectively by reference numeral 502), are (1) split off (step 504) from the series of current profiles (in the illustrated example, the "current profile total" and current profile M+l), and (2) become part of the collection of historical profiles 506 (in this example, current profile total minus current profile M). Again, other varia- tions can be employed, and are within the scope of the present invention. Having described the illustrated embodiments of the present invention, it will be apparent that modifications can be made without departing from the spirit and scope of the invention, as defined by the appended claims.

Claims

1. In a content delivery system including a profiling process operable to generate user profiles based on user actions, a method of determining whether a current user of the content delivery system is a previous user of the content delivery system, the method comprising: during operation of the content delivery system by a current user, receiving from the profiling process at least one historical profile representative of a pattern of user actions by at least one respective previous user; during operation of the content delivery system by the current user, receiving from the profiling process a series of most-recent profiles developed over groupings of most-recent user actions of the current user; iteratively comparing, to the at least first historical profile, ones of the series of most-recent profiles; for each comparison, calculating a degree of similarity value repre- sentative of a degree of similarity between a most-recent profile and the historical profile; determining, on the basis of the calculated degrees of similarity, whether the current user is a previous user of the content delivery system.

2. The method of claim 1 wherein the receiving includes receiving multiple historical profiles representative of at least first and second previous users, respectively; and the determining includes determining whether the current user is any of the first or second previous users.

3. The method of claim 1 wherein the determining includes desig- nating a degree of similarity threshold value, and wherein if a comparison between a most-recent profile and a historical profile results in a degree of similarity value less than the threshold value, then the most-recent profile is used to begin generating a new historical profile.

4. The method of claim 1 wherein the groupings are groupings of 2ⁿ most-recent user actions

5. The method of claim 1 wherein the iterative comparison utilizes decaying averages of the most-recent profiles

6. The method of claim 1 wherein the user actions include remote control button pushes.

7. The method of claim 1 wherein the user actions include digital clicks in a clickstream.

8. The method of claim 1 wherein the content viewing system is a television content viewing system.

9. The method of claim 1 wherein the content viewing system is an Internet-based content viewing system.

10. The method of claim 1 wherein the content viewing system is operable to enable viewing content from multiple sources including sources selected from the group consisting of television and the Internet.

11. In a content delivery system including a profiling process operable to generate user profiles based on user actions, a system operable to determine whether a current user of the content delivery system is a previous user of the content delivery system, the system comprising: means operable during operation of the content delivery system by a current user to receive from the profiling process at least one historical profile representative of a pattern of user actions by at least one respective previous user; means operable during operation of the content delivery system by the current user to receive from the profiling process a series of most- recent profiles developed over groupings of most-recent user actions of the current user; means for iteratively comparing, to the at least first historical profile, ones of the series of most-recent profiles; means for calculating, for each comparison, a degree of similarity value representative of a degree of similarity between a most-recent profile and the historical profile; and means for determining, on the basis of the calculated degrees of similarity, whether the current user is a previous user of the content delivery system.

12. The system of claim 11 wherein the receiving includes receiving multiple historical profiles represen- tative of at least first and second previous users, respectively; and the determining includes determining whether the current user is any of the first or second previous users.

13. The system of claim 11 wherein the determining includes designating a degree of similarity threshold value, and wherein if a comparison between a most-recent profile and a historical profile results in a degree of similarity value less than the threshold value, then the most-recent profile is used to begin generating a new historical profile.

14. The system of claim 11 wherein the groupings are groupings of 2ⁿ most-recent user actions

15. The system of claim 11 wherein the iterative comparison utilizes decaying averages of the most-recent profiles

16. The system of claim 11 wherein the user actions include remote control button pushes.

17. The system of claim 11 wherein the user actions include digital clicks in a clickstream.

18. The system of claim 11 wherein the content viewing system is a television content viewing system.

19. The system of claim 11 wherein the content viewing system is an Internet-based content viewing system.

20. The system of claim 11 wherein the content viewing system is operable to enable viewing content from multiple sources including sources selected from the group consisting of television and the Internet.