GB2545708A - Matching media content with associated metadata - Google Patents

Abstract

A system is disclosed for identifying equivalent or similar content media, such as for example music tracks, published across different content providers and streaming services accessible on networks such as the Internet. The approach utilizes metadata that may differ from different content providers associated with the same media content. Probabilistic classification methods and machine learning techniques are utilized to compute a probability of two pieces of media content such as music tracks being the same music track from identifying and matching the metadata of the media content files without inspecting the media content of the media content files themselves.

Description

MATCHING MEDIA CONTENT WITH ASSOCIATED METADATA FIELD OF THE INVENTION

This invention relates generally to a method and system for matching media content with associated metadata, and more particularly to identifying and matching music audio files with the same music or audio content with different associated metadata from different content providers.

BACKGROUND OF THE INVENTION

The recent emergence of multiple media content providers and streaming services accessible on networks such as the Internet has resulted in a significant fragmentation of the on-demand media content market such as the music access market. Users can potentially use one or more of the available services to access the same piece of media content, such as music tracks, in various ways. That presents a significant problem for services which require identification of a unique piece of media content to perform certain tasks such as identifying and retrieving particular media content from a particular provider. For example, if a user posts a music track onto, for example a social media application service, from one content provider for other users, the other users may only be able to access the music track if they subscribe to the same content provider as the user sharing the track. Other users subscribed to other content providers are denied access to the posted music track.

The technical problem of identifying equivalent pieces of media content across different media content providers and services is compounded by each of the content providers using different ways of identifying music internally, often assigning their own identifiers rather than using international standards such as International Standard Recording Code (ISRC). Moreover the metadata associated with the media content such as a music track is often stored in proprietary format and lacks certain pieces of information allowing for easy and exact identification of the content.

There have been attempts to identify equivalent or similar media content from different media content providers. One such method for identifying music tracks is the so-called process of music track fingerprinting. This technique is either restricted to calculating a simple hash checksum of a content file or in its more advanced form involves computing the representative features of the sound file such as Mel Frequency Cepstral Coefficient (MFCC), pitch frequencies, Fourier transform coefficients as well as their variance and derivatives. A carefully chosen representation ensures that each song is represented uniquely enabling a comparison and retrieval of files in a relatively low dimensionality space compared to the original files.

However fingerprinting requires pre-processing of each and every file to compute the representation - an operation that is computationally expensive and time consuming. Moreover it requires an access to all the resources of all the content providers - an assumption that cannot be reasonably made.

Another approach attempted involves matching based on the metadata associated with each piece of content. Such metadata for media content, such as music tracks, commonly contains information about the artist, track title, album name, and more. The associated metadata is different for each of the content providers, often differing not only in schema but also in the specific pieces of information stored together with the content. For example some content providers provide information about track title, artist, and album as separate, normalised entries in the metadata record, while other content providers allows only retrieval of free text title and description with no guarantee on their content and actual format.

Such match service and cross-provider playback implement rule based systems consists merely of a list of rules embodied in a type of knowledge base, and an inference engine for performing an action based on interaction of the input and rule base. Such rule based approaches have limitations, for example, they allow only for a discretised, for example binary, decision whether the songs are equivalent or not, and they lack any notion of confidence or fuzzy reasoning unless expressed explicitly by the rules. Another limitation is the creation and maintenance of the rules and knowledge base which relies on large amount of labelled training data and human expertise and domain specific knowledge and experience often resulting in number of heuristics or beliefs implemented within the inference engine which results in high maintenance overhead.

There is a need for a method and apparatus for identifying or matching media content with the same media content across different media content providers and services that address or at least alleviate some of the problems and/or limitations discussed above.

SUMMARY OF THE INVENTION A first aspect of the invention is a method of matching media files with associated metadata, comprising receiving a match criteria for retrieving a media content file with associated metadata; searching a plurality of media content files with associated metadata from at least one content provider; and retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.

In an embodiment the match criteria is a metadata of a source media content file. The probability distribution may be based on a supervised or unsupervised training.

In an embodiment the probability distribution is based on Bayes' theorem.

An aspect of the invention is an apparatus of matching media files with associated metadata, comprising a searching module for receiving a match criteria for retrieving a media content file with associated metadata, and searching a plurality of media content files with associated metadata from at least one content provider; and a matching module for retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.

In an embodiment the apparatus further comprises a training module and a prediction module. The training module may comprise a supervised or unsupervised training element. The training module may comprise a learning unit. The prediction module may comprise a classifier for providing a result of match or not match in view of the search criteria.

An aspect of the invention is a computer implemented method of matching media files with associated metadata, the method comprises receiving a match criteria for retrieving a media content file with associated metadata; and at a processor, searching a plurality of media content files with associated metadata from at least one content provider; and retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.

An aspect of the invention is a user electronic device for matching media files with associated metadata, the device comprises a memory storing machine readable instructions; and a processor configured to execute the machine readable instructions to implement the steps of the method according to the first aspect of the invention.

An aspect of the invention is a system of matching media files with associated metadata, the system comprises a server having a memory for storing machine readable instructions and a processor configured to execute the machine readable instructions; a first user electronic device having a memory for storing machine readable instructions and a processor configured to execute the machine readable instructions; the server and the first user electronic device being configured to communicate with each other over a network; wherein the server and the first user electronic device interoperate to implement the steps of the method according to the first aspect of the invention.

An aspect of the invention is a computer readable medium storing machine readable instructions executable by a processor of a user electronic device for implementing the steps of the method according to the first aspect of the invention.

An aspect of the invention is a computer readable medium storing machine readable instructions executable by a processor of a server for implementing the steps of the method according to the first aspect of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings incorporated herein and forming a part of the specification illustrate several aspects of the present invention and, together with the description, serve to explain the principles of the invention. While the invention will be described in connection with certain embodiments, there is no intent to limit the invention to those embodiments described. On the contrary, the intent is to cover all alternatives, modifications and equivalents as included within the scope of the invention as defined by the appended claims. In the drawings: FIG. 1 shows a schematic block diagram of a system for searching, comparing and matching audio file songs from different content providers in accordance with an embodiment of the invention; FIG. 2 shows a schematic block diagram of a processor of FIG. 1 shown in more detail for matching similar audio files with different associated metadata in accordance with an embodiment of the invention; FIG. 3 shows a schematic diagram of an audio file with metadata in accordance with an embodiment of the invention; FIG. 4 shows a schematic block diagram of a learning unit and classifier with supervised training and prediction in searching and matching modules for matching audio file metadata in accordance with an embodiment of the invention; FIG. 5 shows a schematic block diagram of a learning unit and classifier with unsupervised training and prediction in searching and matching modules for matching audio file metadata in accordance with an embodiment of the invention; and FIG. 6 is a flow chart of a method of matching similar audio files with different associated metadata in accordance with an embodiment of the invention.

DETAILED DESCRIPTION

The following description is of preferred embodiments by way of example only and without limitation to the combination of features necessary for carrying the invention into effect. A method and system is disclosed for identifying equivalent or similar content media such as for example music tracks published across different content providers and streaming services are accessible on networks such as the Internet. The approach utilizes metadata that may differ from different content providers associated with the same media content. Probabilistic classification methods and machine learning techniques are utilized to compute a probability of two pieces of media content such as music tracks being the same music track from identifying and matching the metadata of the media content files without inspecting the media content of the media content files themselves.

As described herein a media file is any type of file such as audio, sound, music, tracks, visual, graphic, photographic, textual, or the like. In the illustrative description and embodiments described herein, the content media is music or audio content files. Flowever, it will be appreciated that the media content may take different forms in addition to or in place of music or audio content files in accordance with embodiments of the invention.

Referring to FIG. 1, a schematic block diagram 10 is shown of a system for searching, comparing and matching audio file songs in accordance with an embodiment of the invention. The system 10 shows a user device 12 with a music content file 14 such as a music track with associated metadata. In one embodiment, the user device is a mobile telephone. Flowever, the user device 10 may be any device such as a MP3 player, a lap top computer, a personal digital assistant (PDA), tablet, or the like which is provided with a communication interface and music playing capabilities. The associated metadata of the music track may take a particular format as defined by a particular content provider. The user device 12 posts or shares the file 14 with a file post module 16 to other users and user devices 18 having a file access module 20. The user devices 12,18 may communicate with each other via a server 24 either directly, as shown by dashed arrows, or indirectly via a network such as the internet 22 or the like. The server has a memory 26 for storing information and applications and a processor 28 for executing the applications and processing of different modules of the applications, such as social media application services, and the like. The processor is also able to access a number of similar or identical music content files 30,32,34 that may be provided by a number of content providers 31,33,35 such as first content provider 31, second content provider 33, nth or third content provider 35, or the like, where n is the number of different content providers. It will be appreciated that any number of music content files 14 may be accessed from a single content provider 24, individually from respective different content providers, or multiple content providers with at least one music content file. The music content file 14 is shown in FIG. 1 to be resident on the share user device 12, however, it will be appreciated that the music content file 14 may be resident remote from the share user device 14, elsewhere with the network, such as on the server 24, a content provider 31,33,35 or the like. Once the same matching music content files are retrieved from the different content providers, each having different associated metadata formats respective of the different content providers, the access user 18 is provided with the content file having the metadata in the format of the content provider that the access user is able to access.

The matching content files may already be resident on the access user's device that the access had downloaded or accessed previously. The access user 18 may subscribe to any number of music content providers, and may therefore be able to access any number of music content files having different metadata formats from any number of subscribed content providers. FIG. 2 shows a schematic block diagram 40 of a processor 28 of the server 24 of FIG. 1 shown in more detail for matching similar audio files 14 with the user audio file 14 or request by comparing the metadata having a first set of associated metadata or search criteria that may differ from another audio file with the same or similar audio content having a second set of associated metadata in accordance with an embodiment of the invention. The processor 22 of the server 16 may comprise a user interface module 42 for providing an interface with the share user device 12 and access user device 18, a searching module 44 for searching and identifying identical or similar music content files 14 by inspecting the associated metadata of the content files 14, and a matching module 46 for comparing and selecting the most probable or most likely match. The interaction of the different modules 42,44,46 is described in more detail with reference to FIG. 4-6. FIG. 3 shows a schematic diagram 60 of an audio file 14 with metadata in accordance with an embodiment of the invention. The audio file 14 comprises audio file data 62 comprising the music content, and audio file metadata 64 comprising at least one metadata feature such as first metadata feature 66, second metadata feature 68, and the like. FIG. 4 shows a schematic block diagram 100 of a learning unit 110 and classifier 120 with supervised training elements 111 and prediction elements 121 in searching 44 and matching 46 modules for matching audio file metadata in accordance with an embodiment of the invention. The metadata input 102 is received at an extractor 104 for extracting the features 106 of the metadata as shown as first set of features 108, second feature 109, and the like. The learning unit 110 receives the features 106 and a label 112 for supervised training. The learning unit 110 provides a classifier 120 with a result of the training. The classifier 120 then receives input 122 for prediction with details of another audio file for matching. The input 122 is received and extracted at feature extractor 104 into a second set of features 126 comprising at least one feature. The classifier then determines the label 128 based on the results of the training and the supervised prediction. The label 112 may be for example, Title A and Title B match (0,1). The features may be, for example, titles, artist, album, ISRC, title description, artist description, album description, duration, version, or the like. FIG. 5 shows a schematic block diagram 150 of a learning unit 160 and classifier 170 with unsupervised training 151 and prediction 171 in searching 44 and matching 46 modules for matching audio file metadata in accordance with an embodiment of the invention. The metadata input 152 is received at an extractor 154 for extracting the features 156 of the metadata as shown as first set of features 158, second feature 159, and the like. The learning unit 160 receives the features 156 for unsupervised training. In unsupervised training there is no label 112 as described in the embodiment with reference to FIG. 4. The learning unit 160 provides a classifier 170 with a result of the training. The classifier 170 then receives input 172 for prediction with details of another audio file for matching. The input 172 is received and extracted at feature extractor 174 into a second set of features 176 comprising at least one feature. The classifier then determines the label 178 based on the results of the training and the unsupervised prediction.

An embodiment of the invention employs a probabilistic classifier to perform the matching. A probabilistic classifier predicts, given a sample input, a probability distribution over set of possible outcomes, rather than just the most likely outcome associated with the considered input. This enables a classification with a degree of certainty providing additional information aiding further decision process and action execution. A classifier is a function that assigns a class label fto a sample x: f = fix) (1) where sample x comes from some set X of possible inputs, and y is a class label from finite set of possible outcomes Y defined prior to the training. In this case the possible inputs x are metadata document pairs. In case of song matching it is simply a binary set: match; not-match. Probabilistic classifiers generalise this notion by assigning probabilities to all y E Y given x EX. Being a probability distribution P(Y\X) these probabilities sum up to one. The hard decision can be done using an optimal decision rule which is simply choosing the most probable outcome as:

(2)

Several classifiers including naive Bayes and logistic regression are naturally probabilistic. For others such as support vector machines methods exist to turn them into probabilistic classifiers. The main difference between using probabilistic classifiers and rule based systems is lack of need for producing and maintaining knowledge base. This has a significant impact on the performance of the system. Using machine learning techniques enables creation of systems that can learn association between sample input and desired outcome automatically, choose the most relevant features and infer their relative contribution as well as account for missing data without need for human experts. Moreover the models can be easily extended and retrained to account for new features, user feedback and insights gained during their operation.

At the same time access to the probability distribution over possible outcomes allows for more complicated action of the output, such as sorting matches based on the match probability allowing for the most likely matches to be used while providing a natural fail-back mechanism, e.g., in case of geographical restrictions or availability. In addition, direct access the probability of each of the possible outcomes enables a further decision such as rejecting matches based on weak evidence, i.e., with probability P(Y = $\X) falling below predefined threshold τ.

An implementation of an embodiment is based on a variation of naive Bayes classifier, computing the probability distribution P(Y\X) where X is set of features computed on a pair of compared metadata documents such as binary flags indicating title, artist and album name match, multinomial features expressing degree of match between various bits of information, or numerical features such as difference in number of words in various fields.

The distribution /’(/ΙΧ) is estimated by applying Bayes' theorem: (3)

In practice the denominator of this equation is effectively constant, while the numerator is an equivalent of joint probability and can be rewritten using chain rule and repeated application of conditional probability definition as:

(4) (5) (6) (7) followed by introducing string independence assumption between the features:

(8) (9) (10)

Resulting in the conditional distribution over Y expressed as:

(11)

Where Z = P(x) dependent on xl,... xn only and constant if the values of feature variables are known. The decision rule is therefore expressed as:

(12) where priors P(y) and likelihoods P(xi\y) can be feasibly estimated from training data.

In an embodiment the classifier may be based on a maximum likelihood (ML) probability such as achieved with a Naive Bayes type classifier 120 in the system of FIG. 4, with a training module and a prediction module. The aim is achieved with a match between the target song and search result across all services or content providers. The highest match confidence/relevance from each content provider is selected and retrieved, or at least the link to access the song. The training module estimates the probabilities of each feature based on their frequencies in training data. The prediction module outputs the probability of each class based on unseen data, such as in equation 11, wherein y = Ck. Accordingly, with this configuration, the system requires labels, and needs to have the truths set. The features of the classifier are based on assumptions. A supervised embodiment as shown in FIG. 4 requires a knowledge of output class, and data is labelled with a class or value, where truth is available, where the goal is to predict class and value. An unsupervised embodiment as shown in FIG. 5 requires no knowledge of output class, and data is unlabelled or the values are unknown, with a goal of finding patterns and grouping.

In an embodiment, the classifier is unsupervised, and a bag of word approach is implemented with a relevance score, such as shown in FIG. 5. Such relevance score may be shown as:

(13) (14)

The relevance score is based on term frequency, inverse document frequency, average document length, and the like. In this embodiment, relevance score is not strictly a probability, and computed relevance scores are not probabilities. In an embodiment a ranking function or weighting scheme such as best match or matching (BM) ranking functions such as BM25 or BM25F, is arranged to behave like a probabilistic classifier. For example a distribution of relevance scores for match and mismatch, and then compute a probability of a match given the BM25 relevance score P(Y = y\score). For simplicity, the scores may be assumed to be normally distributed. In this configuration the prediction can improve based on user feedback, and can learn on the whole dataset. Certain fields can be weighted fields, such as important fields of artist, title, version, and the like.

In relevance feedback of the unsupervised approach , estimates the probability that the word (or term) belongs to a relevant or non-relevant document. In the supervised approach, such as the naive Bayes, can retrain and/or update model based on feedback, and ranking may retrieve relevant tracks, then re-rank based on relevance feedback. A self-learning aspect of an embodiment of the invention comprises a monitor match quality and self-learning algorithm. A matching lifecycle comprises: capture user data, process data, update quality statistics, update matching algorithm, update future matches, and display best match. In data collection, any track mismatch and correction is acknowledged such as with CM-1202. The track mismatch and correction may be recorded such as: 1) track appeared on screen (with metadata), track was clicked, track was reported incorrect, next track in the order of relevance was selected. In such a self-learning and match quality checking, potential list events may comprise: open player, play clicked, play started, rewind and fast forward, ad played, pause, stop, close player, heartbeat (still playing), next song played (automatic), manual skip, or the like. The relevance may be inferred from implicit information such as for example: user played some percentage of song, user does not report song, user does not swipe to next song, user explicitly confirms relevance, or the like. In addition, the statistics on match may provide insight on the relevance, for example, what proportion was correct, per user and per search at each position, the number of tacks matched correctly, number of tracks reported of manual curation of tracks, and how likely that an incorrect match is reported, or the like. FIG. 6 is a flow chart of a method 200 of matching similar audio files with different associated metadata in accordance with an embodiment of the invention. The match criteria or user file request is received 202. A plurality of media content files, each media content file having an associated metadata, is searched by the metadata and accessed 204 from at least one content provider. A media content file is selected 206 over a probability distribution of match criteria with associated metadata being most probable match. The selection of the media content file with the most probable match of the most likely to have media content identical or similar to the user file request may be achieved with a training module and prediction module. The training module may be supervised or unsupervised.

Embodiments of the invention have been described herein, including the best mode known to the inventors for carrying out the invention. Variations of those preferred embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the invention to be practiced otherwise than as specifically described herein. Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by the applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the invention unless otherwise indicated herein or otherwise clearly contradicted by context.

The method and system in accordance with an embodiment of the invention is for identifying and retrieving equivalent or similar content media such as for example music tracks published across different content providers and streaming services are accessible on networks such as the Internet. The actual files may be accessible by the access user directly or via a link, or the like. The approach utilizes metadata that may differ from different content providers associated with the same media content. The probabilistic classification methods and machine learning techniques are utilized to compute a probability of two pieces of media content such as music tracks being the same music track from identifying and matching the metadata of the media content files without inspecting the media content of the media content files themselves. The method provides a cross-content provider match of shared tracks, and provides the best match to the access user. If a share user posts a track to share, the method may first search the database and file stores for matching content already downloaded or resident on the access user's device.

The methods and apparatus described may be implemented at least in part in software. Those skilled in the art will appreciate that the apparatus described may be implemented using general purpose computers or using bespoke equipment. Those skilled in the art will appreciate that the foregoing has described what is considered to be the best mode and, where appropriate, other modes of performing the invention, the invention should not be limited to the specific configurations and methods disclosed in this description of an embodiment of the invention. Those skilled in the art will recognise that the invention has a broad range of applications, and that the embodiments may take a wide range of modifications without departing from the inventive concept as defined by the appended claims.

The hardware elements, operating systems and programming languages of such computers are conventional in nature, and it is presumed that those skilled in the art are adequately familiar therewith. Of course, the server functions may be implemented in a distributed fashion on a number of similar platforms, to distribute the processing load.

Hence, aspects of the methods and apparatus described herein can be executed on a mobile station and on a computing device such as a server. Program aspects of the technology may be thought of as "products" or "articles of manufacture" typically in the form of executable code and/or associated data that is carried on or embodied in a type of machine readable medium. "Storage" type media include any or all of the memory of the mobile stations, computers, processors or the like, or associated modules hereof, such as various semiconductor memories, tape drives, disk drives and the like, which may provide storage at any time for the software programming. All or portions of the software may at times be communicated through the Internet or various other telecommunication networks. Such communications, for example, may enable loading of the software from one computer or processor into another computer or processor. Thus, another type of media that may bear the software elements includes optical, electrical and electromagnetic waves, such as used across physical interfaces between local devices, through wired and optical landline networks and over various air-links. The physical elements that carry such waves, such as wired or wireless links, optical links or the like, also may be considered as media bearing the software. As used herein, unless restricted to tangible non-transistory "storage" media, terms such as computer or machine "readable medium" refer to any medium that participates in providing instructions to a processor for execution.

Hence, a machine readable medium may take many forms, including but not limited to, a tangible storage medium, a carrier wave medium or physical transmission medium. Non-volatile storage media include, for example, optical or magnetic disks, such as any of the storage devices in any computer(s) or the like, such as may be used to implement the data aggregator, the customer communication systems, etc. shown in the drawings. Volatile storage media include dynamic memory, such as main memory of such a computer platform. Tangible transmission media include coaxial cables; copper wire and fiber optics, including the wires that comprise a bus within a computer system. Carrier-wave transmission media can take the form of electric or electromagnetic signals, or acoustic or light waves such as those generated during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media therefore include for example: a floppy disk, a flexible disk, hard disk, magnetic tape, any other magnetic medium, a CD-ROM, DVD OR DVD-ROM, any other optical medium, punch cards paper tape, an other physical storage medium with patterns of holes, a RAM, a PROM and EPROM, a FLASH-EPROM, any other memory chip or cartridge, a carrier wave transporting data or instructions, cables or links transporting such a carrier wave, or any other medium from which a computer can read programming code and/or data. Many of these forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to a processor for execution.

While the foregoing has described what are considered to be the best mode and/or other examples, it is understood that various modifications may be made therein and that the subject matter disclosed herein may be implemented in various forms and examples, and that the teachings may be applied in numerous applications, only some of which have been described herein. It is intended by the following claims to claim any and all applications, modifications and variations that fall within the true scope of the present teachings. Variations of those preferred embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the invention to be practiced otherwise than as specifically described herein. Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by the applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the invention unless otherwise indicated herein or otherwise clearly contradicted by context.

Reference in this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Moreover, various features are described which may be exhibited by some embodiments and not by others. Similarly, various requirements are described which may be requirements for some embodiments but not other embodiments.

It should be understood that the elements shown in the figures, may be implemented in various forms of hardware, software or combinations thereof. Preferably, these elements are implemented in a combination of hardware and software on one or more appropriately programmed general-purpose devices, which may include a processor, memory and input/output interfaces.

The present description illustrates the principles of the present invention. It will thus be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described or shown herein, embody the principles of the invention and are included within its spirit and scope.

Moreover, all statements herein reciting principles, aspects, and embodiments of the invention, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Additionally, it is intended that such equivalents include both currently known equivalents as well as equivalents developed in the future, i.e., any elements developed that perform the same function, regardless of structure.

In the claims hereof, any element expressed as a means for performing a specified function is intended to encompass any way of performing that function including, for example, a) a combination of circuit elements that performs that function or b) software in any form, including, therefore, firmware, microcode or the like, combined with appropriate circuitry for executing that software to perform the function. The invention as defined by such claims resides in the fact that the functionalities provided by the various recited means are combined and brought together in the manner which the claims call for. It is thus regarded that any means that can provide those functionalities are equivalent to those shown herein.

While the invention has been illustrated and described in detail in the drawings and foregoing description, the same is to be considered as illustrative and not restrictive in character, it being understood that only exemplary embodiments have been shown and described and do not limit the scope of the invention in any manner. It can be appreciated that any of the features described herein may be used with any embodiment. The illustrative embodiments are not exclusive of each other or of other embodiments not recited herein. Accordingly, the invention also provides embodiments that comprise combinations of one or more of the illustrative embodiments described above. Modifications and variations of the invention as herein set forth can be made without departing from the spirit and scope thereof, and, therefore, only such limitations should be imposed as are indicated by the appended claims.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.

It is to be understood that, if any prior art publication is referred to herein, such reference does not constitute an admission that the publication forms a part of the common general knowledge in the art.

Claims (19)

1. CLAIMS:

   1. A method of matching media files with associated metadata, comprising: receiving a match criteria for retrieving a media content file with associated metadata; searching a plurality of media content files with associated metadata from at least one content provider; and retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.
2. 2. The method of claim 1 wherein the match criteria is the metadata of a source media content file.
3. 3. The method of claim 1 or 2 wherein the probability distribution is based on a supervised training.
4. 4. The method of claim 1 or 2 wherein the probability distribution is based on an unsupervised training.
5. 5. The method of any one of the preceding claims wherein the probability distribution is estimated based on Bayes' theorem:
6. 6. The method of claim 5 wherein the probability distribution is based on P(y\X) being: P(y)p(xl\y)P(x2\y,xi) ...p(xn\y,xl,x2, ...xn— 1)
7. 7. The method of claim 6 wherein the probability distribution is a conditional distribution over Y being:

   where Z = P(x) dependent on xl,... xn.
8. 8. The method of claim 7 wherein the probability distribution, a decision rule is:

   wherein P(y) priors and P(xi|y)likelihood is estimated from training data.
9. 9. An apparatus of matching media files with associated metadata, comprising: a searching module for receiving a match criteria for retrieving a media content file with associated metadata, and searching a plurality of media content files with associated metadata from at least one content provider; and a matching module for retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.
10. 10. The apparatus of claim 9 further comprising a training module and a prediction module.
11. 11. The apparatus of claim 10 wherein the training module comprises a supervised training element.
12. 12. The apparatus of claim 10 wherein the training module is unsupervised.
13. 13. The apparatus of any one of claims 10-12 wherein the training module comprises a learning unit.
14. 14. The apparatus of any one of claims 10-12 wherein the prediction module comprises a classifier for providing a result of match or not match with the search criteria.
15. 15. A computer implemented method of matching media files with associated metadata, the method comprising: receiving a match criteria for retrieving a media content file with associated metadata; and at a processor, searching a plurality of media content files with associated metadata from at least one content provider; and retrieving a selected media content with associated metadata within the plurality of media content files wherein the selected media content is selected over a probability distribution of the match criteria with the associated metadata being most probable match.
16. 16. A user electronic device for matching media files with associated metadata, the device comprising: a memory storing machine readable instructions; and a processor configured to execute the machine readable instructions to implement the steps of the method according to any one of claims 1 to 8.
17. 17. A system of matching media files with associated metadata, the system comprising: a server having a memory for storing machine readable instructions and a processor configured to execute the machine readable instructions; a first user electronic device having a memory for storing machine readable instructions and a processor configured to execute the machine readable instructions; the server and the first user electronic device being configured to communicate with each other over a network; wherein the server and the first user electronic device interoperate to implement the steps of the method according to any one of claims 1 to 8.
18. 18. A computer readable medium storing machine readable instructions executable by a processor of a user electronic device for implementing the steps of the method according to any one of claims 1 to 8.
19. 19. A computer readable medium storing machine readable instructions executable by a processor of a server for implementing the steps of the method according to any one of claims 1 to 8.