Brand and Consumer Protection

From SAM
Jump to: navigation, search

The term Brand Protection refers to the right of owners to protect their brands and reputation from unauthorized use (copyright). Consumer Protection is a safeguard towards customers regarding inappropriate content. In other words, consumer should be protected from being harmed by unexpected content such as random, non-verified information or comments coming from prosumers. This is essential when the consumers are children but also for social network spam in general.


Sort text classifier.JPG
In the past, the delivery of content through the Internet and media platforms was always performed with respect to the protection of the content owners and their brand as well as the safety of the content audience. The respective processes were quite simple and straightforward in the era where web was considered as static. The roles of the various stakeholders were absolutely clear and hence the requirements for Brand and Consumer Protection could be easily addressed. However, in the era of Web 2.0 and now the era of social media, the processes for creation and delivery of content have introduced enormous complexity due to an increasing number of sources and destinations of the content.

Today content is created and delivered literally everywhere and the boundaries of the actors responsible for the ownership, the control and the curation of content are not always clear. New terms have been introduced, such as the term “prosumer” which examplars the ambiguities in this unsolidified environment. Thus, the requirements for effective practices to deal with Brand and Consumer Protection have been considerably increased.

In order to create a safe environment, security mechanisms and legal boundaries are needed. The security mechanisms ensure that the content created or distributed is safe (not malicious, inappropriate). On the other hand, legal boundaries ensure the correct and legal distribution of the content in order to avoid piracy and illegal use of the content .In this article, an analysis of the mechanisms and legal aspects is taking place.

Relevance to SAM

Brand and Consumer Protection is a broad term and many technical solutions, tools and methodologies can be found in the literature as well as products that deal with the Web, the Internet and Social Media. In order to better identify the most relevant approaches that could be the baseline for the Brand and Consumer Protection in SAM, it is necessary to analyse the SAM user scenarios and the related user stories. To this direction the production and prosumption user scenarios were examined and the requirements related to this domain were identified.

In the introduction section, two terms were mentioned: security mechanisms and legal boundaries. The question is why SAM needs these two elements in order to create a safe environment-platform.

SAM Platform interfere with the social networks; more specifically, content from social networks is published in SAM Platform. This attribute creates the necessity for mechanisms that evaluate and categorise the content in order to avoid inappropriate language, spam content and malicious content published in SAM Platform through social networks.

SAM Platform also interfere with content providers who provide content. This content is accessible through the Sam Platform. However, the distribution of these content must be accommodated with legal aspects about the content, more specifically aspects about the ownership of the content, in which circumstances the content can be used and other legal rules. In collusion, legal aspects about the usage and the distribution of the content must be created in order to avoid and eliminate privacy violations or content piracy. The Brand and Consumer Protection in the context of SAM is described in the T5.4 Brand and consumer protection in WP5 Content Syndication and Delivery.

State of the Art Analysis

Brand Protection Systems

The consideration of brand protection systems in this subsection focuses on protection systems for digital media and systems. Brand protection systems and mechanisms for physical goods (e.g. anti-counterfeiting tags on products) are not considered as they are not relevant for SAM.

Generally, brand protection systems as they are of concern for SAM specialise in the identification and cataloguing of suspected brand infringements, usually at the level of the World Wide Web and/or at the level of specific popular platforms such as YouTube or Facebook. Integrated brand protection systems that are offered to individual companies – both as services and products – combine content analysis mechanisms and/or social network analysis mechanisms with report generation features and interfaces to third-party services in cases where automated action for brand protection purposes can be taken (e.g. to YouTube or to Facebook). The table below lists and characterises a few commercial brand protection system propositions that are available at the time of writing. Due to a lack of available reliable market data, it is difficult to assess market shares for the specific systems; the listed systems should be considered as representative examples.



NetNames services


Provides a mixture of consulting services and real-time monitoring of websites and public social media in order to identify copyright infringements



Provides brand usage infringement detection and monitoring tool as a service to businesses



Provides brand online reputational risk identification and management services to businesses

Figure 1: Representative Examples of Brand Protection Systems

In addition to such brand protection systems, providers of online platforms that significantly rely on user-generated content may implement systems for preventing copyright infringements, for instance in order to identify unauthorised usage of music in uploaded video clips. The techniques applied in this type of deployment are fundamentally the same as for brand protection systems and have been outlined in the previous subsections. Commercial platform operators frequently also employ takedown services, where copyright and brand owners can file complaints in response to which operators take down content described as infringing copyright. Such services are commonly operated in a fully automated fashion and allow individual companies to submit the output of their brand protection systems to platform operators.

User Profiling and Protection

User profiling is an essential ingredient for a protected system. Especially Social Networks have established mechanisms for user Profiling-Modeling. The Amount of the users that use the social Networks Create the necessity for such mechanisms. In order to have clear view of what user profile is we provide a definition for user profile. The profile of a user u ∈ U is a set of weighted concepts where with respect to the given user u for a concept c ∈ C its weight w(u, c) is computed by a certain function w.

               P(u) = {(c, w(u, c))|c ∈ C, u ∈ U}

Here, C and U denote the set of concepts and users respectively. [1]

There are several mechanisms and Systems that embed user profiling in them in order to achieve their goals or to maximize efficiency.

Adaptive Web Search Based on User Profile Constructed

The World Wide Web (WWW) is growing day by day exponentially. This effect makes the search and the classification of the content very difficult and chaotic. The search engines have created a decentralized way of finding useful information for every user by adapting the search results through user profiling. This means the same search query returns different results depending on the information that is gathered for the user.  Experimental results show that search system create user profiling based on modified collaborative filtering with detailed analysis of user's browsing history in one day[2].

Automatic Web User Profiling and Personalization Using Robust Fuzzy Relational Clustering

This framework creates user profiling using server access logs based on robust fuzzy relational clustering. The clustering process creates robust profiles. The profiles are computed from prior traversal patterns of the users on the web sites they use[3].

Cross-system user modeling and personalization on the Social Web

Cross-system user modeling uses distributed form-based and tag-based user profiles. The social networks provide a very large data set in order to accomplice accurate user profiling. The modeling analyzes completeness, replication and consistency of the social network profiles (form-based profiles). The form based analysis is not enough because these profiles may be sparse. The tag-based profiling creates a more accurate profile of a user[4]

Ontological User Profiling

Ontological user profiling is used for recommending systems, at the moment is used within the recommender for academic research papers. The system creates a user profile from inconspicuously monitored behavior and relevance feedback, representing the profiles in terms of a research paper topic ontology[5].

Web user behavioral profiling for user identification

This approach creates a user profile by monitoring web browsing behavior. This process captures the strength of users’ behavioral patterns in order to make the profiles[6].

Content Classification and Filtering

Sort Text Classification

The SAM is inextricably connected with the social networks. A fundamental issue that comes with this attribute is the ability of the user to control the content that is posted on his account in order to avoid unwanted material. There are many systems available for content filtering, although giving the nature of the content that is created/published on social networks, this systems will not work properly(i.e sort facebook posts, small tweets). Taking this under consideration a need to use a different type of filtering is created. Such systems are the sort text classifiers who use specific algorithms to detect the nature of the content in small texts[7].

Tools, Frameworks and Services

There are several mechanisms for creating user profiles that can be embedded in SAM.

Compression Framework for Generating User Profiles

The user profiles are generated by extracting information and features of the past online behavior of the users. More specifically this framework depends on finding compressed representation of the behavior of the user by selecting dominant features contributing to the Kullback-Leibler divergence between the default distribution over user actions and users specific properties.[8] This Framework uses the profiles that creates mainly for recommendation systems, but a strong user profile can be a great contribution to the safety of a System SAM tends to create.

Framework for Flexible User Profile Mashups

Within this framework a user modeling framework is created which is called Grapple User Modeling Framework(GUMF). This framework is created in order to allow heterogeneous Systems to connect to this framework and collect user profiles, for that reason it provides a flexible user model format which allows new types of statements and derivation rules[9]. One of the main attributes of the SAM system is the connection of heterogeneous systems (i.e. HbbTV, e-commerce sites, Social Networks) this attribute creates the necessity of a flexible user profiles , created by collected and processed data from several systems.

Twitter-Based User Modeling Framework

Twitter-Based User Modeling Framework uses semantic enrichment on twitter messages in order to create specific entities and give semantic value to Twitter activities. Additionally this framework links the twitter messages to external resources with various strategies and techniques. With the combination of the semantic enrichment and the external linkage of the messages users create the framework provides a strong and rich user profiles. SAM system has a strong linkage with the Social Networks[10]. In order to obtain a complete and strong user profile SAM needs to have social aspect attributes in the profiles that creates.

Related Projects


The RADICAL project (Rapid Deployment for Intelligent Cities and Living) will enable the development and deployment of smart-city services through a novel platform by leveraging emerging Social Networks technologies, Internet‐of‐Things (IoT) infrastructures and the R&D results from the SmartSantander, BonFIRE, SocIoS, and +Spaces projects, while at the same time handling legal, governance and socio‐economic issues[11].

Standards and Policies

A complete analysis of the legal aspects can be found here: Legal_Aspects

SAM Approach

The BCP component is responsible for addressing aspects related to Asset content by applying protection filters in order to avoid unexpected content affecting both brands and End Users integrity. Next paragraph shows the different subcomponents of the BCP, the logical connections that have been established between them and the relations with other components and actors in the SAM platform.

Architecture and Dependencies

This component, consisting both from GUI elements and services, as presented in the following image, provides several key features to both brands and End Users in order to apply conent protection filters with the purpose of avoiding that unexpected content could affect their integrity. More specifically, the BCP component:

  • It applies filters to address brand protection mechanism; owners should have the possibility to protect their brands from unauthorised use.
  • Theses filters will serve to address inappropriate content filtering; consumer should be protected also from being harmed by unexpected content (e.g., access to adult content should only be possible in acceptable way).
  • SAM Platform Manager uses the BCP Rules Editor in order to create and modify all those necessary rules that the SAM Platform needs to filter the inappropriate content.


Implementation and Technologies

After Extended Analysis and comparison the most appropriate technologies for the frontend and the backend have been selected.

Frontend Technologies (User Interface)

The user interface BCP Rules Editor has been implemented using the technology AngularJS. Also because the aforementioned interface is embedded inside the marketplace, it is reasonable to choose this technology to provide a common look and feel[12].

Backend Technologies (Web Services)

The BCP prototype uses the JAX-RS technologies and more specifically the Jersey framework[13] in order to implement the RESTful Web Services for this component. In addition, the BCP API uses the Swagger framework[14] to obtain an interactive documentation. This should considerably ease the implementation, deployment and testing of the BCP-backend environment.


A summary of the tasks carried out for each subcomponent during the first and second versions of the prototype is shown in the following table.

Subcomponent Task
BCP Rules Manager Define JSON format for the Assets, user profiles and rules (including possible fields, conditions and values)
BCP Rules Manager Create, update, remove and load rules functions from a JSON file
Content Filter Implement Asset filtering given an owner and a user profile
BCP Rules Manager

Content Filter

Expose the functionalities of both components as RESTful Services
BCP Rules Editor Implement create, update, remove and load rules functionalities using the BCP Rules Manager RESTful Services
BCP Rules Editor Include SAM look and feel
BCP Rules Editor Integrate the interface in the Administration Tool
BCP Rules Manager Update the JSON format for complex rules explicitly specifying logical operators
BCP Rules Manager Functions to create, read, update and delete rules from the Cloud Storage
Content Filter Update the evaluation of rules to integrate the new complex rules format
BCP Rules Editor Create, read, update and delete functionalities following the new complex rules format
BCP Rules Editor Differentiate fields belonging to Asset Structure or User Profile

Functionality and UI Elements

This section introduces the required steps to create a rule that checks the state of an Asset, both in the BCP Rules Editor and in the BCP Rules Manager subcomponent. Finally, the Content Filter subcomponent is presented.

Creating a Rule through the BCP Rules Editor

In order to access the BCP Rules Editor, the user needs to be registered in the SAM platform and logged in the Administration Tool or the Marketplace. The “Brand and Consumer” option in the left side menu provides access to the BCP Rules Editor (see figure below).


In the Rules Editor interface, the button “New” opens the “New Rule” dialog shown in the figure above. In this window, users can add a description for the new rule, select the owner of the rule[15] , activate it, and insert different filtering criteria.
With respect to filtering criteria, rules can contain different statements (new statements can be added by means of the “+Criterion” button) combined with logical operators (“+AND” and “+OR” buttons). First, all logical operators must be inserted, and then one or more statements can be added nested for each operator.
As can be seen in the figure below, each statement contains the following elements:

  • The first column represents the subject of the statement, indicating which End User or Asset feature is going to be evaluated in the rule. In this second prototype, these features are clearly differentiated to indicate which ones belong to the Asset Structure and which ones refer to the User Profile. Possible values in this column include “User Age” and “Rating”, indicating the age of the user and the rating of an Asset (i.e. indicating suitable ages for a film) respectively.
  • The second column shows the conditions that can be applied to the selected field (e.g. “GREATERTHAN” and “EQUALS”)
  • The last column represents the value with which the field is going to be compared by means of the condition selected. It can be a numeric value (e.g. “18” for the field “User Age”) or a string (e.g. “Movie” for the field “Asset Type”).


The previous figure shows a rule in the user interface composed of three criteria combined by two logical operators (an “OR” and an “AND”). The first statement establishes that the User Age is less than “18”. The second one indicates that the “Rating” of the Asset is “18” (suitable only for adults). The last one specifies that the “User Country” is “DE” (Germany). The first and the second criteria are connected using the “AND” logical operator, whereas the latter is combined to them using the “OR” logical operator. This rule could be described as “Filter Assets classified for adults to underage users from Germany”.

Creating a Rule through the BCP Rules Manager

The interface previously shown relies on the BCP Rules Manager subcomponent to manage rules. The documentation demo page of the BCP Rules Manager (see figure below) provides an interface where back-end functionalities corresponding to this subcomponent can be tested. This documentation exposes for each HTTP method at least one input example, its response codes, and the URL of the RESTful Services to try the subcomponent out.

In order to create a new rule, the method requires a JSON object containing the description of the rule, its owner, a flag indicating if it is active or not, and the filtering criteria. The rule “Filter Assets classified for adults to underage users from Germany” coded in JSON format is shown below.

  "description": "Filter adult content to underage users from Germany",
  "owner": "DW",
  "active": true,
  "id": 3,
  "criteria": {
    "type": "OR",
    "items": [
        "type": "AND",
        "items": [
            "type": "SIMPLE",
            "criterion": {
              "field": {
                "iname": "Rating",
                "jname": "asset.rating",
                "jtype": "int"
              "condition": "GREATERTHANOREQUALS",
              "value": "18"
            "type": "SIMPLE",
            "criterion": {
              "field": {
                "iname": "User Age",
                "jname": "user.age",
                "jtype": "int"
              "condition": "LESSTHAN",
              "value": "18"
        "type": "SIMPLE",
        "criterion": {
          "field": {
            "iname": "User Country",
            "jname": "",
            "jtype": "java.lang.String"
          "condition": "EQUALS",
          "value": "DE"

Figure beneath shows a fragment of the documentation demo page for the “Create a new rule” operation. In the “Parameter” section, the body text box can be filled with a JSON example (as the one previously shown) and then check the result using the “Try it out!” button. The section “Response Messages” provides information on the possible responses of the BCP Rules Manager when a rule is created.

Filtering an Asset

The Content Filter subcomponent filters Assets on demand given a target End User and a rule owner. The back-end functionalities of this subcomponent can be tested using the documentation demo page mentioned in the previous section (see figure below). This documentation exposes for this HTTP method an input example, its response codes, and the URL for the RESTful Service of this subcomponent.

Similarly to what was explained in the previous section, figure beneath shows a fragment of the documentation demo page for the “Filters content” operation. The input of this component is a JSON object containing a list of Assets to filter, the profile of the target End User who wants to consume these Assets, and the owner of the Assets (whose rules must be applied). As a result, this operation returns the same list of Assets marked as approved or not according to the owner rules and the User Profile.


Latest Developments

During the last period of the SAM life cycle some new developments have been made on the Brand and Consumer component. These changes have been mostly focused on: using the last SAM Asset structure; and implement a Black and White list mechanism, which indicates whether an Asset Provider can link her content to the content of another provider (white list) or not (black list).

These progresses obviously affected the User Interface. On the one hand, the Rules Editor has been adapted to deal with the creation and edition of rules for the new enriched SAM Assets. On the other hand, a new User Interface to manage the rules Black and White list mechanism has been developed, which can be seen below.

Similarly, the back-end was also affected. The BCP Rules Manager has not modified the methods or its input/outputs. However, the rules have been enhanced to use the last SAM Asset structure and to allow complex rules, that is, to compare the value of two fields. A new back-end interface was created to deal with Black and White lists, as shown below.
BCP BlackwhitelistBackend.png

Lastly, the Content Filter subcomponent was adapted to use the new assets and rules, as well as the Black and White lists restrictions. NewContentFilterBackend.jpg


  1. Sugiyama, Kazunari, Kenji Hatano, and Masatoshi Yoshikawa. "Adaptive web search based on user profile constructed without any effort from users." Proceedings of the 13th international conference on World Wide Web. ACM, 2004.
  2. Adaptive Web Search Based on User Profile
  3. Automatic Web User Profiling and Personalization Using Robust Fuzzy Relational Clustering
  4. Cross-system User Modeling and Personalization on the Social Web
  5. Web Search Personalization with Ontological User Profiles
  6. Web user behavioral profiling for user identification
  7. A System to Filter Unwanted Messages from OSN User Walls
  8. Shi, Xiaoxiao, et al. "A compression framework for generating user profiles."Workshop of the 33 rd Annual International. 2010.
  9. Abel, Fabian, et al. "A framework for flexible user profile mashups." (2009).
  10. Abel, Fabian, et al. "Analyzing user modeling on twitter for personalized news recommendations." User Modeling, Adaption and Personalization. Springer Berlin Heidelberg, 2011. 1-12.
  15. In the current prototype, rules are not attached to a particular Protection Manager in the SAM platform, since profiling functionalities are not yet available. Thus, a dropdown menu allows selecting an owner for the sake of testing the component