US 20130110502A1 (19) United States (12) Patent Application Publication (10) Pub. No.: US 2013/0110502 A1 Issa et al. (43) Pub. Date: May 2, 2013

(54) SYSTEM AND METHOD FOR INTERNET Publication Classi?cation RADIO STATION PROGRAM DISCOVERY (5 1) Int. Cl. (71) Applicant: Lemi Technology, LLC, Wilmington, G10L 15/18 (2006.01) DE (US) G10L 15/26 (2006.01) (52) US. Cl. (72) Inventorsl Alfredo C- Issa, Apex, NC (US); CPC ...... G10L 15/18 (2013.01); G10L 15/265 Richard J. Walsh, Raleigh, NC (US); (201301) glsristopher M- Amidon, Apex, NC USPC ...... 704/9; 704/231; 704/235 (57) ABSTRACT (73) Assignee: LEMI TECHNOLOGY, LLC, . . . . . - - An Internet rad1o station program discovery servlce is pro W1lm1ngton, DE (US) V1ded.. A plurahty. of Internet rad1o. station. programs 1s. (21) Appl NO _ 13/716 507 obtained. For each Internet radio station program of the plu ' " ’ rality of Internet radio station programs, the Internet radio - _ station program is dynamically categorized by mapping a (22) Flled' Dec‘ 17’ 2012 dynamically identi?ed topic of the Internet radio station pro . . gram to a content classi?er. A User is enabled to discover an Related U's' Apphcatlon Data Internet radio station program of interest from the plurality of (63) Continuation of application No. 12/273,709, ?led on Internet radio station programs based on the dynamic catego Nov. 19, 2008, noW Pat. No. 8,359,192. riZations for the plurality of Internet radio station programs. /10

INTERNET INTERNET RADIO RADIO STATION STATION PROGRAM PROVIDER(S) DISCOVERY SERVICE M E

USER DEVICE USER DEVICE USER DEVICE @ m 16-N

INTERNET RADIO INTERNET RADIO . . . INTERNET RADIO PLAYBACK FUNCTION PLAYBACK FUNCTION PLAYBACK FUNCTION

20-1 20-2 20-N Patent Application Publication May 2, 2013 Sheet 1 0f 10 US 2013/0110502 A1 Patent Application Publication May 2, 2013 Sheet 2 0f 10 US 2013/0110502 A1

SUBSCRIBE TO AND/OR DOWNLOAD A NUMBER OF INTERNET RADIO STATION PROGRAMS

II FOR EACH RADIO STATION PROGRAM, ANALYZE AUDIO CONTENT OF THE RADIO STATION PROGRAM TO DYNAMICALLY CATEGORIZE THE RADIO STATION PROGRAM OVER TIME

II ENABLE USERS TO DISCOVER /-104 RADIO STATION PROGRAMS OF INTEREST BASED ON THE DYNAMIC CATEGORIZATIONS OF THE RADIO STATION PROGRAMS

FIG. 2 Patent Application Publication May 2, 2013 Sheet 3 0f 10 US 2013/0110502 A1

PERFORM SPEECH-TO-TEXT CONVERSION AND NATURAL LANGUAGE PROCESSING ON THE AUDIO CONTENT <— OF THE RADIO STATION PROGRAM TO IDENTIFY TOPIC(S) I MAP TOPIC(S) TO AN ONTOLOGY TO IDENTIFYA NUMBER OF CONTENT CLASSIFIERS IN THE ONTOLOGY ASSOCIATED WITH THE TOPIC(S) I SCORE THE IDENTIFIED CONTENT CLASSIFIERS

206 MORE AUDIO YES CONTENT?

FIG. 3 Patent Application Publication May 2, 2013 Sheet 4 0f 10 US 2013/0110502 A1

/24

DAN BILL MARINO PARCELLS 26-4 26-6

FIG. 4 Patent Application Publication May 2, 2013 Sheet 5 0f 10 US 2013/0110502 A1

MIAMI DON INDIANAPOLIS DOLPHINS SHULA COLTS 26-1 26-2 26-3 SCORE: 20 SCORE: 10 SCORE: 2

DAN BILL N EW YO RK MARINO PARCELLS J ETS 26-4 26-6 26-9 SCORE: 15 SCORE: 0 SCORE: 0

NEW ENGLAND DALLAS PATRIOTS COWBOYS 26-5 26-8 N EW YO RK SCORE: 3 GIANTS SCORE: 0 26-7 SCORE: 0

MARINO OF FAME TOPICS SHULAi MARINOi RADIO STATION | > PROGRAM STREAM I

FIG. 5 Patent Application Publication May 2, 2013 Sheet 6 0f 10 US 2013/0110502 A1

|V||A|V|| DON INDIANAP LI DOLPHINS SHULA COLTSO 8 26-1 26-2 m SCORE: 8 SCORE: 2 SCORE: O

DAN BILL NEW YORK MARINO PARCELLS JETS 26-4 26-6 26-9 SCORE: 3 SCORE: 15 SCORE: 5

NEW ENGLAND DALLAS PATRIOTS COWBOYS 26-5 26-8 NEW YORK SCORE: 23 26-7 SCORE: 17

LESS INFLUENCE MORE INFLUENCE ’ T ‘. l I ' I HALL : MARINO OF FAME JETS GIANTS . I l l I l I TOPICS SHULAl MARINOi PARCELLSI l GIANTSl GIANTSI RADIO I i I i > STATION l l I TIME PROGRAM TIME TIME TIME STREAM 0 x Y

FIG. 6 Patent Application Publication May 2, 2013 Sheet 7 0f 10 US 2013/0110502 A1

28 COWBOYS 26-8

PATRIOTS M FIG.7 DOLPHINS 26-1 DAN MARINO 26-4 [3O ENFL E|College EIFootball ‘HE... E|Sports ‘E... Patent Application Publication May 2, 2013 Sheet 8 0f 10 US 2013/0110502 A1

§\ 8K v22E66wh_<9m 56mo>526$2, 0562Ema .2: .2: mg

ow m Patent Application Publication May 2, 2013 Sheet 9 0f 10 US 2013/0110502 A1 |§\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

8\|rme; %\|Nwe; EU625E:w: 850% EQ25 _W_05:22 Patent Application Publication May 2, 2013 Sheet 10 0f 10 US 2013/0110502 A1 E:\

Tmm .GE:.

2929232200wo/Egw.565 wo?mwkz6595a % 6E2. US 2013/0110502 A1 May 2,2013

SYSTEM AND METHOD FOR INTERNET aspects of the invention, and together With the description RADIO STATION PROGRAM DISCOVERY serve to explain the principles of the invention. [0008] FIG. 1 illustrates a system including an Internet RELATED APPLICATIONS radio station program discovery service according to one [0001] This application is a continuation of US. patent embodiment of the present invention; application Ser. No. 12/273,709 (now US. Pat. No. ), [0009] FIG. 2 is a How chart illustrating operation of the titled “SYSTEM AND METHOD FOR INTERNET RADIO Internet radio station program discovery service of FIG. 1 STATION PROGRAM DISCOVERY”, ?led on Nov. 19, according to one embodiment of the present invention; 2008, the entire disclosure of Which is here incorporated by [0010] FIG. 3 is a How chart illustrating a process for reference. dynamically categoriZing an Internet radio station program according to one embodiment of the present invention; FIELD [0011] FIG. 4 illustrates an exemplary ontology of content classi?ers; [0002] The present invention relates to the classi?cation or [0012] FIGS. 5 and 6 graphically illustrate dynamic iden categorization of Internet radio station programs. ti?cation of topics of an Internet radio station program and BACKGROUND mapping of the topics to the ontology of FIG. 4 according to one embodiment of the present invention; [0003] Through portals such as blogtalkradio.com and [0013] FIGS. 7 and 8 are exemplary Graphical User Inter various other distribution sources, there are noW thousands of faces (GUIs) through Which a user is enabled to discover Internet radio talk shoWs. Further, projections shoW that the Internet radio stations of interest according to one embodi number of Internet radio talk shoWs is expected to quickly ment of the preset invention; rise. One issue resulting from the large number of Internet [0014] FIG. 9 is an exemplary GUI of an Internet radio radio talk shoWs available is that users need a Way to quickly station playback function of a user device of a user including and easily discover Internet radio talk shoWs of interest. an index or list of topics discussed in an Internet radio station program being played according to one embodiment of the SUMMARY present invention; [0004] The present invention relates to an Internet radio [0015] FIG. 10 is a block diagram of a server hosting the station program discovery service. In general, the discovery Internet radio station program discovery service of FIG. 1 service subscribes to or doWnloads a number of Internet radio according to one embodiment of the present invention; and station programs. For each radio station program, the discov [0016] FIG. 11 is a block diagram ofone ofthe user devices ery service analyZes audio content of the radio station pro of FIG. 1 according to one embodiment of the present inven gram to dynamically categoriZe the radio station program tion. over time. The discovery service also enables users to dis cover radio station programs of interest based on the dynamic DETAILED DESCRIPTION categoriZations of the radio station programs. The users may [0017] The embodiments set forth beloW represent the nec be enabled to search for radio station programs of interest essary information to enable those skilled in the art to practice based on the dynamic categoriZations of the radio station the invention and illustrate the best mode of practicing the programs, navigate or broWse the radio station programs invention. Upon reading the folloWing description in light of based on the dynamic categoriZations of the radio station the accompanying draWing ?gures, those skilled in the art programs, or the like. Will understand the concepts of the invention and Will recog [0005] In one embodiment, in order to dynamically catego niZe applications of these concepts not particularly addressed riZe each of the radio station programs, the discovery service herein. It should be understood that these concepts and appli performs speech-to-text conversion and natural language pro cations fall Within the scope of the disclosure and the accom cessing on the audio content of the radio station program to panying claims. identify one or more topics of the radio station program. The [0018] FIG. 1 illustrates a system 10 incorporating an Inter discovery service then maps the one or more topics to an ontology or similar taxonomy describing relationships net radio station program discovery service 12 (hereinafter betWeen a number of content classi?ers in order to identify “discovery service 12”) according to one embodiment of the present invention. As illustrated, the system 10 includes the one or more content classi?ers associated With the one or more topics of the radio station program. The identi?ed con discovery service 12, one or more Internet radio station pro viders 14, and a number of user devices 16-1 through 16-N tent classi?ers may then be scored. This process is repeated to continually process the audio content of the radio station connected via a netWork 18. The netWork 18 may be any type program to dynamically categoriZe the radio station program. of Wide Area NetWork (WAN), Local Area NetWork (LAN), or any combination thereof. Further, the netWork 18 may [0006] Those skilled in the art Will appreciate the scope of the present invention and realiZe additional aspects thereof include Wired components, Wireless components, or both after reading the folloWing detailed description of the pre Wired and Wireless components. As an example, the netWork 18 may be a global netWork such as the Internet Where each of ferred embodiments in association With the accompanying the discovery service 12, the one or more Internet radio sta draWing ?gures. tion providers 14, and the user devices 16-1 through 16-N BRIEF DESCRIPTION OF THE DRAWING have access to the netWork 18 via a Wired connection, a local FIGURES Wireless connection (e.g., IEEE 802.11x, Bluetooth, or the like), or a mobile telecommunications netWork connection [0007] The accompanying draWing ?gures incorporated in (e.g., Wideband Code Division MultipleAccess (W-CMDA), and forming a part of this speci?cation illustrate several Global System for Mobile Communications (GSM), or the US 2013/0110502 A1 May 2,2013

like). Lastly, the user devices 16-1 through 16-N have asso 12 may or may not complete analysis of the radio station ciated users 20-1 through 20-N. programs before the users 20-1 through 20-N are enabled to [0019] As discussed below, the discovery service 12 oper discover radio station programs of interest. In one embodi ates to subscribe to or download a number of Internet radio ment, using the user 20-1 as an example, the Internet radio station programs (hereinafter “radio station programs”) pro playback function 22-1 of the user device 1 6-1 may enable the vided by the one or more Internet radio station providers 14. user 20-1 to enter one or more search terms. In response, the For example, the radio station programs may be talk shows. Internet radio playback function 22-1 may query the discov The one or more Internet radio station providers 14 may be ery service 12 for radio station programs having content servers operated and maintained by one or more commercial classi?ers matching the one or more search terms. In another entities such that the corresponding Internet radio stations are embodiment, the user 20-1 may interact with the discovery commercial radio stations, servers operated by individual service 12 via the Internet radio playback function 22-1 to users such that the corresponding Internet radio stations are navigate or browse a catalog or index of content classi?ers. personal radio stations, or the like. The discovery service 12 Then, upon selecting a desired content classi?er, the user 20-1 then dynamically categoriZes the radio station programs over may be presented with a list of radio station programs asso time. In addition, the discovery service 12 enables the users ciated with the desired content classi?er. 20-1 through 20-N to discover radio station programs of [0024] FIG. 3 is a ?ow chart illustrating a process for interest based on the dynamic categoriZations of the radio dynamically categoriZing one of the radio station programs of station programs. FIG. 2 according to one embodiment of the present invention. [0020] Each ofthe user devices 16-1 through 16-N may be First, the discovery service 12 performs speech-to-text con any type of user device enabled to provide playback of Inter version and natural language processing on audio content of net radio station content such as, for example, a personal the radio station program to identify one or more topics dis computer, a mobile smart phone, a portable media player cussed in the radio station program (step 200). Note that as having network capabilities, or the like. The user devices 16-1 used herein speech-to-text conversion includes, but is not through 16-N include Internet radio playback functions 22-1 limited to, traditional speech-to-text conversion via word rec through 22-N, respectively. The Internet radio playback func ognition, speech-to-text conversion by ?rst converting speech tions 22-1 through 22-N may be implemented in software, into phonemes using a technique such as that used by Nexidia hardware, or a combination thereof. The Internet radio play and then converting the phonemes into text, or the like. Also back functions 22-1 through 22-N enable the users 20-1 note that if the radio station program is received via a stream through 20-N to interact with the discovery service 12 in ing audio channel, the discovery service 12 may buffer the order to discover radio station programs of interest to the audio content of the radio station program and perform users 20-1 through 20-N. In addition, the Internet radio play speech-to-text conversion and natural language processing on back functions 22-1 through 22-N enable the users 20-1 the audio content of the buffered radio station program. In one through 20-N to listen to desired radio station programs dis embodiment, bookmarks or markers may be created and covered via the discovery service 12. stored for the identi?ed topics. The bookmarks may be imple [0021] FIG. 2 is a ?ow chart illustrating operation of the mented as time offsets from the beginning of the radio station discovery service 12 according to one embodiment of the program. Alternatively, if the audio content of the radio sta present invention. First, the discovery service 12 subscribes to tion program is buffered by the discovery service 12, the or downloads a number of radio station programs from the bookmarks may be implemented as pointers or references to one or more Internet radio station providers 14 (step 100). In locations within a buffer storing the buffered audio content. one embodiment, the Internet radio stations, and thus the [0025] Next, the discovery service 12 maps the one or more radio station programs provided on the Internet radio stations, topics to an ontology or similar taxonomy of content classi are received by the discovery service 12 via corresponding ?ers to identify a number of content classi?ers associated streaming audio channels. In another embodiment, the radio with the one or more topics (step 202). More speci?cally, in station programs are downloaded by the discovery service 12 one embodiment, each content classi?er in the ontology is as, for example, podcasts. In yet another embodiment, some associated with one or more keywords and/or descriptions. radio station programs may be provided via corresponding Then, for a particular topic, the discovery service 12 identi?es streaming audio channels while other radio station programs the content classi?ers in the ontology that have the topic listed may be downloaded as, for example, podcasts. as an associated keyword and/ or for which the topic is [0022] For each radio station program, the discovery ser included in the corresponding descriptions. vice 12 analyZes audio content of the radio station program to [0026] In one embodiment, an ontology mapping rights dynamically categoriZe the radio station program over time (OMR) ?le may be provided to the discovery service 12 in (step 102). As discussed below, in the preferred embodiment, association with one or more of the radio station programs. In the discovery service 12 analyZes the audio content of the general, the OMR ?le for a radio station program may assist radio station program using speech-to-text conversion and the discovery service 12 in the classi?cation of the radio natural language processing. However, in an alternative station program and/or restrict classi?cation of the radio sta embodiment, tags identifying topics of the radio station pro tion program. More speci?cally, the OMR ?le for a radio gram over time may be included in or provided in association station program may include information explicitly identify with the radio station program. The tags may be inserted by, ing permissible content classi?ers from the ontology to which for example, the provider, by a community of users that have the topics identi?ed for the radio station program may be listened to the radio station program, or the like. mapped, information identifying non-permissible content [0023] In addition, the discovery service 12 enables the classi?ers from the ontology to which the topics for the radio users 20-1 through 20-N to discover radio station programs of station program may not be mapped, or both. interest based on the dynamic categoriZations of the radio [0027] The OMR ?le may also include one or more sample station programs (step 104). Note that the discovery service voice clips to assist the discovery service 12 in the speech US 2013/0110502 A1 May 2,2013

to-text and natural language processing. The OMR ?le may knowledge such as Wikipedia. Using Wikipedia as an also include information identifying one or more topics dis example, each Wikipedia page represents an ontological node cussed in the radio station program and times at which the where a title of the page is used as the corresponding content topics are discussed and/or information identifying refer classi?er. Therefore, each of the content classi?ers 26-1 ences to content classi?ers in the ontology and times at which through 26-9 may correspond to a Wikipedia page. The Wiki the content classi?ers are applicable to the radio station pro pedia pages are analyZed to discover relationships between gram. This may be done as an alternative to or in addition to the pages, and thus relationships between the content classi the speech-to-text and natural language processing and map ?ers. For example, each Wikipedia page, and thus the corre ping done by the discovery service 12. The OMR ?le may also sponding content classi?er, may be said to be related to all include demographic information describing target users for other Wikipedia pages that the Wikipedia page references, or the radio station program. The OMR ?le may include one or to which the Wikipedia page has a link. Further, each Wiki more buffer rules de?ning how long the discovery service 12 pedia page may be analyZed to identify one or more keywords may buffer the audio content of the radio station program. associated with the corresponding content classi?er and/or Additionally, the OMR ?le may include a link to where an facts regarding the corresponding content classi?er. Still fur archive of the radio station program will be stored by the ther, on Wikipedia, each page is classi?ed into one or more corresponding Internet radio station provider 14. This link hierarchical categories. As such, each Wikipedia page may be may be provided to and used by the Internet radio playback analyZed to identify the categories for the corresponding con functions 22-1 through 22-N if the users 20-1 through 20-N tent classi?er. Fur‘ther, a hierarchical list or index of catego desire to listen to the radio station program after the radio ries may be obtained directly from Wikipedia or based on an station program has aired. Lastly, the OMR ?le may include analysis of the Wikipedia pages. metadata about the radio station program such as, for example, information about the radio station program, infor [0030] For this example, the ontology 24 includes the fol mation about guests on the radio station program, or the like. lowing information for each of the content classi?ers 26-1 This metadata may assist the discovery service 12 in the through 26-9. identi?cation of topics and/or the mapping of topics to the ontology. TABLE 1 [0028] In this embodiment, once the content classi?ers Content Classi?er Dolphins associated with the one or more topics are identi?ed, the identi?ed content classi?ers are scored (step 204). For each of Relationships Don Shula Bill Parcells the identi?ed content classi?ers, the score of the content Dan Marino classi?er may be a function of a relevance assigned to the one or more topics mapped to the content classi?er. The relevance Keywords Miami of a topic may be determined during speech-to-text conver Football National Football League sion and natural language processing based on, for example, NFL a density or frequency of that topic in the audio content of the Super Bowl VI radio station program, strength of the topic, speech metrics Dan Marino associated with the use of the topic or terms related to the Don Shula Fins topic in the audio content of the radio station program. As an example, speech metrics may be a volume of a speaker’s Facts Lost Super Bowl VI to Dallas voice, which may be described as shouting, yelling, normal, Cowboys or whispering. In addition, the content classi?ers may be Perfect season in 1972 Won Super Bowl VII scored as a function of the relative importance of each of the Won Super Bowl VIII content classi?ers in the ontology. The relative importance of Coached by Don Shula from each of the content classi?ers in the ontology may be deter 1970-1995 mined using, for example, PageRank or Eigenvector algo Categories NFL Teams rithms that determine “centrality” of nodes, or content clas Sports in Miami, Florida si?ers, in the ontology. Note that any suitable scoring technique may be used to score the content classi?ers such that the scores of the content classi?ers are indicative of the relevance of the content classi?ers to the radio station pro gram. The exemplary scoring techniques discussed herein are TABLE 2 not intended to limit the scope of the present invention. Content Classi?er Don Shula [0029] FIGS. 4 through 7 graphically illustrate the process Relationships Dolphins of FIGS. 2 and 3 according to an exemplary embodiment of Colts the present invention. FIG. 4 illustrates an exemplary ontol ogy 24, or at least a subset of an exemplary ontology 24, used Keywords Dolphins Head Coach to dynamically classify radio station programs. As illustrated, Football the ontology 24 de?nes relationships between a number of National Football League content classi?ers 26-1 through 26-9, which are represented NFL as nodes in the ontology 24. The ontology 24 may be created Super Bowl VII manually or based on an of?cial ontology such as OpenCyc, Super Bowl VIII Baltimore Colts which is an open source ontology maintained by Cycorp, Inc. Hall of Fame In yet another embodiment, the ontology 24 may be generated by analyZing contents of a publicly available collection of US 2013/0110502 A1 May 2,2013

TABLE 2-c0ntinued TABLE 5

Content Classi?er Don Shula Content Classi?er Patriots

Facts Head Coach of Miami Dolphins Relationships Dan Marino from 1970-1995 Bill Parcells

Won Super Bowl VII . . . Won Super Bowl VIII Keywords Pats Coached Miarni Dolphins to a New England perfect season in 1972 Boston Inducted to Hall of Fame in Football 1997 National Football League

. . . NFL Categories NFL Coaches Super Bowl XX Miarni Dolphins Coaches Super Bowl XXXI Baltimore Colts Coaches Super Bowl XXXVI NFL Hall of Fame Super Bowl XXXVIII Super Bowl XXXIX Torn Brady Bill Parcells

TABLE 3 Facts Won Super Bowls XXXVI, Content Classi?er Colts XXXVIII’ and XXXIX Lost Super Bowls XX and XXXI Relationships D911 Shula Coached by Bill Parcells frorn Keywords Baltimore 1993- 1 99 6

Indianapolis - - - Don Shula Categories NFL Tearns NFL Sports in Boston, Football Massachusetts Peyton Manning

Facts Won ?ve NFL chalnpionships including two Super Bowl titles Won Super Bowl XLI against the Chicago Bears TABLE 6 Coached by Don Shula from 1963-1969 Content Classi?er Bill Parcells Cat?gori?s NFL Twins Relationships Dolphins Patriots Giants Cowboys Jets

TABLE 4 - - - Keywords Football Content Classi?er Dan Marino National Football L‘mgu6 NFL Relationships Dolphins Head Coach Patriots Patriots

. . . Giants Keywords Miarni Cowboys Dolphins Jets Quarterback Executive VP of Football Football Operations National Football League Dolphins

NFL . . . Hall of Fame Facts Coached New York Giants frorn NutriSystern 1983-1990 . . . Coached New England Patriots Facts Quarterback of the Miarni from 1993-1996 Dolphins Coached New York Jets frorn Spokesman for NutriSystern 1997-1999 weight loss prograrn Coached Dallas Cowboys frorn Played against the New 2003-2006 England Patriots on Oct. 11, Executive Vice President of 1996; ?nal score 24-10 Football Operations for Miarni . . . Dolphins frorn 2008-present Categories NFL Quarterbacks . . . Miarni Dolphins Players Categories NFL Head Coaches NFL Hall ofFarne Giants Head Coaches US 2013/0110502 A1 May 2,2013

TABLE 7 because “Don Shula” is a keyword associated with the con tent classi?er 26-1 and/or because “Shula” is mentioned in Content Classi?er Giants one of the facts associated with the content classi?er 26-1. Relationships Bill Parcells Likewise, the topic “Shula” is mapped to the content classi?er 26-2 (“Don Shula”) because “Shula” is found in the title of the Keywords New York content classi?er 26-2. Additionally, the topic “Shula” is Football mapped to the content classi?er 26-3 (“Colts”). In a similar National Football League NFL manner, the topic “Marino” is mapped to the content classi?er Bill Parcells 26-1 (“Dolphins”), the content classi?er 26-4 (“Dan Super Bowl XXI Marino”), and the content classi?er 26-5 (“Patriots”). Lastly, the topic “Hall of Fame” is mapped to the content classi?er Facts NFL team in New York Coached by Bill Parcells from 26-2 (“Don Shula”) and the content classi?er 26-4 (“Dan 1983-1990 Marino”). Categories NFL Tealns [0032] The content classi?ers 26-1, 26-2, 26-3, 26-4, and 26-5 to which the topics “Shula,” “Marino,” and “Hall of Fame” are mapped and scored. In one embodiment, each of the content classi?ers 26-1, 26-2, 26-3, 26-4, and 26-5 is scored based on the relevance of each of the topics that is TABLE 8 mapped to that content classi?er. In addition or alternatively, Content Classi?er Cowboys the content classi?ers 26-1, 26-2, 26-3, 26-4, and 26-5 may be scored based on the degree to which the categories of the Relationships Bill Parcells content classi?ers 26-1, 26-2, 26-3, 26-4, and 26-5 overlap. More speci?cally, using the content classi?er 26-1 as an Keywords Dallas Football example, the content classi?er 26-1 may be scored based on National Football League the degree to which the categories for the content classi?er NFL 26-1 overlap with the categories for the other content classi Bill Parcells ?ers 26-2, 26-3, 26-4, 26-5, and 26-6. Therefore, for example, Facts NFL team in Dallas, Texas if the content classi?er 26-1 is in categoryA (not shown), then Coached by Bill Parcells from the score of the content classi?er 26-1 may be incremented by 2003-2006 +0.1 points for each of the other content classi?ers 26-2, 26-3, 26-4, and 26-5 that is also in category A. Further, if there is Categories NFL Tealns one degree of separation between category A and category B (not shown), then the score of the content classi?er 26-1 may be incremented by +0.05 points for each of the other content classi?ers 26-2, 26-3, 26-4, and 26-5 that is in category B. TABLE 9 Further, if there are two degrees of separation between cat Content Classi?er Jets egory A and category C (not shown), then the score of the content classi?er 26-1 may be incremented by +0.033 points Relationships Bill Parcells for each of the other content classi?ers 26-2, 26-3, 26-4, and 26-5 that is in category C. This may continue for any number Keywords New York Football of degrees of separation. The maximum number of degrees of National Football League separation that is considered may be system de?ned or user NFL con?gurable. Bill Parcells [0033] Next, as additional audio content of the radio station Facts NFL team in New York program is processed and analyzed by the discovery service Coached by Bill Parcells from 12, additional topics are identi?ed and mapped to the ontol 1997-1999 ogy 24. More speci?cally, as illustrated in FIG. 6, from TIME Categories NFL Tealns X to TIME Y, the corresponding audio content of the radio station program is analyzed to identify the topics “Parcells,” “Jets,” and “Giants.” The discovery service 12 then maps the identi?ed topics to the ontology 24 in order to identify one or [0031] FIGS. 5 and 6 graphically illustrate the dynamic more of the content classi?ers 26-1 through 26-9 that are categorization of an exemplary radio station program accord associated with the identi?ed topics. The scores for the iden ing to one embodiment of the present invention. FIG. 5 illus ti?ed content classi?ers are then updated. Note that the topics trates the identi?cation of topics during a ?rst time segment of identi?ed in TIME 0 through TIME X may still in?uence the the audio content of the radio station program, the mapping of scores for the content classi?ers but be given less in?uence or the identi?ed topics to the ontology 24, and the scoring of the less weight than the topics identi?ed in TIME X to TIME Y. associated content classi?ers in the ontology 24 according to [0034] FIG. 7 illustrates an exemplary Graphical User one embodiment of the present invention. More speci?cally, Interface (GUI) 28 enabling a user, such as for example the during a ?rst time segment of the radio station program from user 20-1, to interact with the discovery service 12 to discover TIME 0 to TIME X, the audio content of the radio station radio station programs of interest according to one embodi program is analyzed by the discovery service 12 to identify ment of the present invention. The GUI 28 may be provided the topics “Shula,” “Marino,” and “Hall of Fame.” The topic by the discovery service 12 and presented to the user 20-1 via “Shula” is mapped to the content classi?er 26-1 (“Dolphins”) the Internet radio playback function 22-1 or provided by the US 2013/0110502 A1 May 2,2013

Internet radio playback function 22-1 as an interface to the area 36, the user 20-1 selects the desired radio station pro discovery service 12. The GUI 28 includes a navigation area gram. For example, assume that the user 20-1 selects “Big 30 and an ontology display area 32. The navigation area 30 is Al’s Football talk.” In response, the Internet radio playback optional and may be desirable Where, for example, the ontol function 22-1 of the user device 16-1 of the user 20-1 sends a ogy 24 is large. In this example, the navigation area 30 pre request to the Internet radio station provider 14 for “Big Al’ s sents a hierarchical index or list of categories, and the content Football talk,” and playback begins. Note that a reference to classi?ers in the ontology 24 are categorized into the catego “Big Al’s Football talk,” such as a Uniform Resource Locator ries. In operation, the user 20-1 selects a category of interest, (URL), may be provided to the Internet radio playback func Which in this example is an “NFL” category. Note that the tion 22-1 in advance or may be requested from the discovery NFL category may have a number of sub-categories Which service 12 as needed. In an alternative embodiment, “BigAl’ s may include “NFL Teams,” “NFL Head Coaches,” “NFL Football talk” may be delivered to the Internet radio playback Quarterbacks,” and the like. In response to the user 20-1 function 22-1 of the user 20-1 via the discovery service 12. selecting the NFL category, a ?ltered version of the ontology [0038] In one embodiment, for each radio station program, 24 including only those nodes in the NFL category or one of an index or list of topics discussed in the radio station pro the sub-categories of the NFL category are displayed in the gram is maintained. Further, for each topic, one or more ontology display area 32. The user 20-1 may then select a bookmarks to segments of the radio station program during desired one of the content classi?ers 26-1 through 26-9 in the Which the topic is discussed may also be maintained. FIG. 9 ontology display area 32. illustrates an exemplary GUI 46 of the Internet radio playback [0035] In this example, the user 20-1 selects the content function 22-1 ofthe user device 16-1 ofthe user 20-1 after the classi?er 26-4 (“Dan Marino”). In response, as illustrated in user 20-1 has selected “Big Al’s Football talk” for playback. FIG. 8, a GUI 34 is presented to the user 20-1. The GUI 34 As illustrated, the GUI 46 includes a topic display area 48 includes a radio station program display area 36 and a rela including a list of topics identi?ed for “Big Al’ s Football talk” tionships area 38. The radio station program display area 36 by the discovery service 12. For each topic, a number of presents a list of radio station programs that are associated bookmarks indentify times at Which the topic is discussed. With the Dan Marino content classi?er 26-4. In addition, the Thus, in this example, the topic “Marino” has tWo bookmarks list of radio station programs may be prioritized based on 50 and 52. The bookmark 50 identi?es a time (Time 1) cor relevance to the Dan Marino content classi?er 26-4. In one responding to the beginning of a segment of the radio station embodiment, the relevance of the radio station programs to program during Which the topic “Marino” Was discussed. the Dan Marino content classi?er 26-4 corresponds to the Similarly, the bookmark 52 identi?es a time (Time 2) corre scores for the Dan Marino content classi?er 26-4 generated sponding to the beginning of another segment of the radio by the discovery service 12 for the radio station programs. station program during Which the topic “Marino” Was dis More speci?cally, as discussed above, for each of these radio cussed. By selecting the bookmark 50, playback of “Big Al’ s station programs, one or more topics for the radio station Football talk” jumps to the corresponding time (Time 1) in program Were dynamically identi?ed and mapped to the Dan playback of “Big Al’s Football talk.” LikeWise, by selecting Marino content classi?er 26-4. The resulting scores for the the bookmark 52, playback of “Big Al’s Football talk” jumps Dan Marino content classi?er 26-4 for these radio station to the corresponding time (Time 2) in playback of “Big Al’s programs may then be used to prioritize the radio station Football talk.” programs in the radio station program display area 36. In [0039] FIG. 10 is a block diagram of server 54 hosting the addition or alternatively, the radio station programs in the discovery service 12 according to one embodiment of the radio station program display area 36 may be prioritized present invention. In general, the server 54 includes a control based on Whether or not the radio station programs are cur system 56 having associated memory 58. In this embodiment, rently airing or are live. Still further, the radio station pro the discovery service 12 is implemented in softWare and grams in the radio station program display area 36 may addi stored in the memory 58. HoWever, the present invention is tionally or alternatively be prioritized based on the not limited thereto. The discovery service 12 may be imple demographic information describing target users provided in mented in softWare, hardWare, or a combination thereof. The an OMR ?le for the radio station program as compared to server 54 also includes one or more digital storage devices 60, demographic information describing the user 20-1. As such, at least one communication interface 62 communicatively radio station programs for Which the user 20-1 is a target user coupling the server 54 to the one or more Internet radio station may be given a higher priority in the radio station program providers 14 and the user devices 16-1 through 16-N (FIG. 1), display area 36. Note that other preferences of the user 20-1 and a user interface 64, Which may include components such may additionally or alternatively be used to prioritize the as, for example, a display, one or more user input devices, or radio station programs in the radio station program display the like. area 36. [0040] Note that the server 54 is exemplary. The discovery [0036] The relationships area 38 enables the user 20-1 to service 12 may be implemented on a single server or distrib navigate to nodes in the ontology 24 that are related to the uted over a number of servers. Further, in another embodi current node, Which in this example is the Dan Marino con ment, the discovery service 12 may be distributed over a tent classi?er 26-4. As such, the user 20-1 may navigate to the number of user devices such as, but not limited to, the user Dolphins content classi?er 26-1 by selecting a corresponding devices 16-1 through 16-N. For example, each of the user indicator 40 or navigate to the Patriots content classi?er 26-5 devices 16-1 through 16-N may be responsible for dynami by selecting a corresponding indictor 42. The user 20-1 may cally categorizing one or more radio station programs. return to the GUI 28 (FIG. 7) by selecting a “Top” indicator [0041] FIG. 11 is a block diagram of the user device 16-1 44. according to one embodiment of the present invention. This [0037] If the user 20-1 desires to play one of the radio discussion is equally applicable to the other user devices 16-2 station programs listed in the radio station program display through 16-N. In general, the user device 16-1 includes a US 2013/0110502 A1 May 2,2013

control system 66 having associated memory 68. In this identifying one or more ?rst topics of the Internet radio embodiment, the Internet radio playback function 22-1 is station program based on an analysis of a ?rst portion of implemented in software and stored in the memory 68. HoW audio content of the Internet radio station program cor ever, the present invention is not limited thereto. The Internet responding to a ?rst time segment of the Internet radio radio playback function 22-1 may be implemented in soft station program; Ware, hardWare, or a combination thereof. The user device categorizing the Internet radio station program based on 16-1 also includes a communication interface 70 communi the one or more ?rst topics; catively coupling the user device 16-1 to the netWork 18 (FIG. identifying one or more second topics of the Internet radio 1). Lastly, the user device 16-1 includes a user interface 72, station program based on an analysis of a second portion Which may include a display, one or more user input devices, of the audio content of the Internet radio station program one or more speakers, or the like. corresponding to a second time segment of the Internet [0042] Those skilled in the art Will recognize improve radio station program that is subsequent to the ?rst time ments and modi?cations to the preferred embodiments of the segment of the Internet radio station program; and present invention. All such improvements and modi?cations are considered Within the scope of the concepts disclosed further categorizing the Internet radio station program herein and the claims that folloW. based on the one or more second topics of the Internet radio station program. What is claimed is: 1. A method comprising: 9. The method of claim 8 Wherein dynamically categoriz obtaining a plurality of Internet radio station programs; ing the Internet radio station program comprises, for each for each Internet radio station program of the plurality of additional portion of a plurality of additional portions of the Internet radio station programs, dynamically categoriz audio content of the Internet radio station program corre ing the Internet radio station program by mapping a sponding to a plurality of additional time segments of the dynamically identi?ed topic of the Internet radio station Internet radio station program: program to a content classi?er; and identifying one or more additional topics of the Internet enabling a user to discover an Internet radio station pro radio station program based on an analysis of the addi gram of interest from the plurality of Internet radio sta tional portion of the audio content of the Internet radio tion programs based on the dynamic categorizations for station program; and the plurality of Internet radio station programs. further categorizing the Internet radio station program 2. The method of claim 1 Wherein dynamically categoriz based on the one or more additional topics of the Internet ing the Internet radio station program comprises: radio station program. performing speech-to-text conversion on audio content of 10. The method of claim 1 Wherein obtaining the plurality the Internet radio station program; and of Internet radio station programs comprises, for at least one dynamically categorizing the Internet radio station pro Internet radio station program of the plurality of radio station gram based on results of the speech-to-text conversion. programs, subscribing to a corresponding audio stream. 3. The method of claim 1 Wherein dynamically categoriz 1 1. The method of claim 10 Wherein obtaining the plurality ing the Internet radio station program comprises: of Internet radio station programs further comprises buffering performing speech-to-text conversion and natural lan audio content of each of the plurality of Internet radio station guage processing on audio content of the Internet radio programs, and for each Internet radio station program of the station program; and plurality of Internet radio station programs, dynamically cat dynamically categorizing the Internet radio station pro egorizing the Internet radio station program over time com gram based on results of the speech-to-text conversion prises dynamically categorizing the Internet radio station pro and natural language processing. gram based on the buffered audio content of the Internet radio 4. The method of claim 1 further comprising generating a station program. bookmark for each of the topics. 12. The method of claim 1 Wherein obtaining the plurality 5. The method of claim 1 Wherein dynamically categoriz of Internet radio station programs comprises, for at least one ing the Internet radio station program comprises dynamically Internet radio station program of the plurality of Internet mapping topics to an ontology of content classi?ers to iden radio station programs, doWnloading the at least one Internet tify one or more content classi?ers in the ontology related to radio station program. the topics. 13. A tangible computer readable medium storing a com 6. The method of claim 5 Wherein dynamically categoriz puter program, executable by a machine, for dynamically ing the Internet radio station program based on the topics categorizing Internet radio station programs, the computer further comprises dynamically scoring the one or more con program comprising executable instructions for: tent classi?ers. obtaining a plurality of Internet radio station programs; 7. The method of claim 6 Wherein scoring the one or more content classi?ers comprises scoring each content classi?er for each Internet radio station program of the plurality of of the one or more content classi?ers based on at least one of Internet radio station programs, dynamically categoriz a group consisting of: relevance of one or more of the topics ing the Internet radio station program by mapping a mapped to the content classi?er, relative importance of the dynamically identi?ed topic of the Internet radio station content classi?er in the ontology, and frequency of occur program to a content classi?er; and rence of the one or more topics mapped to the content clas enabling a user to discover an Internet radio station pro si?er. gram of interest from the plurality of Internet radio sta 8. The method of claim 1 Wherein dynamically categoriz tion programs based on the dynamic categorizations for ing the Internet radio station program comprises: the plurality of Internet radio station programs. US 2013/0110502 A1 May 2,2013

14. The computer readable medium of claim 13 Wherein in of the audio content of the Internet radio station program order to dynamically categorize the Internet radio station corresponding to a second time segment of the Internet program, the computer program comprises executable radio station program that is subsequent to the ?rst time instructions for: segment of the Internet radio station program; and performing speech-to-text conversion on audio content of further categorizing the Internet radio station program the Internet radio station program; and based on the one or more second topics of the Internet dynamically categorizing the Internet radio station pro radio station program. gram based on results of the speech-to-text conversion. 18. The computer readable medium of claim 17 Wherein in 15. The computer readable medium of claim 13 Wherein in order to dynamically categorize the Internet radio station order to dynamically categorize the Internet radio station program, the computer program comprises executable program, the computer program comprises executable instructions for, for each additional portion of a plurality of instructions for: additional portions of the audio content of the Internet radio performing speech-to-text conversion and natural lan station program corresponding to a plurality of additional guage processing on audio content of the Internet radio time segments of the Internet radio station program: station program; and identifying one or more additional topics of the Internet dynamically categorizing the Internet radio station pro radio station program based on an analysis of the addi gram based on results of the speech-to-text conversion tional portion of the audio content of the Internet radio and natural language processing. station program; and 16. The computer readable medium of claim 13 Wherein in order to dynamically categorize the Internet radio station further categorizing the Internet radio station program program, the computer program comprises executable based on the one or more additional topics of the Internet instructions for: radio station program. dynamically identifying topics of the Internet radio station 19. A system comprising: program; and at least one communication interface communicatively dynamically categorizing the Internet radio station pro coupling the system to one or more Internet radio station gram based on the topics of the Internet radio station providers and a user device; and program. a control system associated With the at least one commu 17. The computer readable medium of claim 13 Wherein in nication interface and adapted to: order to dynamically categorize the Internet radio station obtain a plurality of Internet radio station programs; program, the computer program comprises executable for each Internet radio station program of the plurality of instructions for: Internet radio station programs, dynamically categorize identifying one or more ?rst topics of the Internet radio the Internet radio station program by mapping a dynami station program based on an analysis of a ?rst portion of cally identi?ed topic of the Internet radio station pro audio content of the Internet radio station program cor gram to a content classi?er; and responding to a ?rst time segment of the Internet radio station program; enable a user to discover an Internet radio station program categorizing the Internet radio station program based on of interest from the plurality of Internet radio station the one or more ?rst topics; programs based on the dynamic catcgorizations for the identifying one or more second topics of the Internet radio plurality of Internet radio station programs. station program based on an analysis of a second portion * * * * *