Scientific Methodology and Work Packages

The project is structured around an empirical cycle that ties together four different streams of work (see the overall project structure picture). Each stream comprises a set of workpackages (WPs). Each workpackage promises deliverables at specified times during the project. Streams run in parallel with dependencies being circular. Significant project achievements, called milestones, indicate points of synchronization or dependence between streams. The empirical cycle by stream and milestone chart below shows mutual dependencies between component tasks in different streams. Dependencies go from left-to-right and downward unless an arrow directs upward. Each arrow represents a data batch. Each M value represents a milestone (see milestones table) and each V number represents a living-labs version deployment (see deliverables table). Numbers in black circles represent data batches (0 to 3). Notice that the first data batch 0, comes from existing data-sets.

Emiric Cycle


Stream 1

Theoretical & algorithmic foundations. Comprises: WP1.1, WP1.2, WP1.3, WP1.4

WP1.1 : Theory of complex techno-social collectives

Work Package 1.1 Leader: ETH Zurich

This workpackage will provide a better theoretical understanding of complex techno-social systems and new suggestions of how to improve them by “guided self-organisation”. This will require both the development of enriched agent-based models and simplified models allowing the formulation of an analytically tractable theory of complex techno-social systems.

Deliverables : D1.1.1, D1.1.2


WP1.2 : Complexity models of trust in networks

Work Package 1.2 Leader: UWAR

This workpackage will deal with an algorithmic specification and comparison of complexity models of agency and communities, studying especially how trust and reputation networks can be formed in techno- social systems and how informational quality can be determined.

Deliverables : D1.2.1


WP1.3 : Theory of quality emergence in complex systems

Work Package 1.3 Leader: UNI Fribourg

The work package will address the question of quality in two complementary ways. First, we will join insights from social psychological models of social influence and the recent achievements in the field of recommender systems to produce plausible models for quality assessment by users. The goal is to understand how quality (as perceived by individuals) emerges. Second, we will develop an agent-based model describing how quality can be estimated.

Deliverables : D1.3.1, D1.3.2


WP1.4 : Modelling the dynamics of quality collectives

Work Package 1.4 Leader: UniS

This workpackage will produce trace-based models, simulations and analysis of the emergence and dynamics of quality collectives.

Deliverables : D1.4.1, D1.4.2


Stream 2

Algorithm design, simulation & evaluation. Comprises: WP2.1, WP2.2, WP2.3

WP2.1 : Algorithms for the emergence of cooperation

Work Package 2.1 Leader: TUD

The work package will design and simulate efficient algorithms for given services and tasks required by the QLectives infrastructure, such as bootstrapping of social structures, cooperative sharing of resources, resource discovery and honest rating and reviewing etc.

Deliverables : D2.1.1, D2.1.2, D2.1.3


WP2.2 : Algorithms for the emergence of collectives

Work Package 2.2 Leader: UNI Fribourg

This work package will produce algorithms that can support the formation of cooperative and efficient social structures.

Deliverables : D2.2.1


WP2.3 : Algorithms for trusted and quality rankings

Work Package 2.3 Leader: USZ

In the QLectives infrastructure, people will generate implicit and explicit ratings of other people and objects regarding trust and quality. If explicit ratings are considered to define weighted links between entities, this will only form a sparse network, where we are given incomplete and also noisy information. The goal of this work package is to develop and analyse algorithms that can extrapolate this information into a state as complete as possible, taking into account results from WP1 and WP3 as well.

Deliverables : D2.3.1


Stream 3

Empirical data-sets collection, processing and validation. Comprises: WP3.1, WP3.2, WP3.3

WP3.1 : Data collection and experimentation

Work Package 3.1 Leader: ETH Zurich

This work package will collect new examples of datasets describing interactions between users of social networking sites, bookmarking sites, P2P networks etc., focusing on tagging behaviour, profile and taste diversity, user activity and social networks. In addition, the project’s own ’Living labs’ (QScience and QMedia) will be instrumented to collect interaction data and data gathered as soon as an initial implementation is piloted with users.

Deliverables : D3.1.1, D3.1.2, D3.1.3


WP3.2 : Data processing and knowledge extraction

Work Package 3.2 Leader: UWAR

This work package will collect very large datasets. In order to make such datasets useful, they need to be cleaned, stored efficiently, and summarised. This will be a computationally intensive process, and may also involve non-automated visual inspection to identify problems in the data or its processing.

Deliverables : D3.2.1, D3.2.2


WP3.3 : Model and algorithm validation and prediction

Work Package 3.3 Leader: CNRS

This work package will test hypotheses of the models developed in Stream 1 and validate algorithms developed in Stream 2. Feedback from this processes will be used by those streams to improve models and algorithms. For this, the large empirical datasets emanating from WP3.1 and WP3.2 will be used in combination with the models and simulations from Stream 1 and 2.

Deliverables : D3.3.1


Stream 4

Platform and living lab implementation. Comprises: WP4.1, WP4.2, WP4.3, WP4.4

WP4.1 : Peer-to-Peer platform - P2P-Qual

Work Package 4.1 Leader: TUD

This work package will create the P2P-Qual platform which combines social networking, facilitation of quality, and scalable P2P technology into a second generation peer-production platform. P2P-Qual will be built on top of the already deployed and mature P2P Tribler code-base. It will implement complexity-science inspired techno-social networking fabric primitives:

  • QFlow: augment P2P transfers of information with audit trial of quality indicators.
  • BWCurrency: Transform P2P bandwidth into a transferable currency.
  • P2PWidgets: run-time deployable code units which extend the platform

Deliverables : D4.1.1, D4.1.2, D4.1.3, D4.1.4


WP4.2 : Scientific innovation living lab - QScience

Work Package 4.2 Leader: UNI Fribourg

This work package will apply algorithms for both first-order (papers, articles and Web resources) and second-order (reviewers, workshops, conferences and journals) rating systems (formulated in stream 1 and stream 2).We will develop and implement proactive social networking algorithms that support the creation / detection of scientific communities allowing like-minded scientists to form links around shared interests and quality assessments.

Deliverables : D4.2.1, D4.2.2, D4.2.3, D4.2.4


WP4.3 : Media distribution living lab - QMedia

Work Package 4.3 Leader: TUD

This work package will provide an alternative to mass media and replace the channel concept with self-organising personalised collectives. It will build upon the P2P-Qual platform and offer seemless download and streaming of open media content, provide social networking functionality allowing users to freely share media with friends and contacts, self-organise, through detection of taste similarity and user rating systems, groups of similar peers, implement social tools allowing users to create a shared viewing experience and promote the discovery of high quality open content by integrating the WP4.4 quality meta-data system.

<


WP4.4 : Quality search and discovery

Work Package 4.4 Leader: IRT

This work package will bring public service broadcasting quality commitment to the Youtube and BitTorrent generation of media distribution, explore ways for sharing quality meta-data from, and with, public broadcasters, give users the ability to rate, moderate, annotate and recommend contents themselves, allow for easy “injection” of existing open meta-data sources, increase quality by reducing spam, incorrect or malicious meta-data via application of novel mechanisms from Stream 1 and 2 and create a global distributed self-organising high quality Electronic Programme Guide (EPG).

Deliverables : D4.4.1, D4.4.2, D4.4.3, D4.4.4



Workpackages that involve dissemination and management activities and are not part of any stream. Comprises: WP5, WP6

WP5 : Dissemination, Collaboration and Exploitation

Work Package 5 Leader: UniS

This work package will produce a functional  project website, deliver dissemination activities, provide reports on Dissemination, Collaboration and Exploitation including evidence of progress on: a) Strategic Impacts; b) dissemination plan and c) a collection of samples of publications, and a collectively authored book.

Deliverables : D5.1, D5.2, D5.3, D5.4, D5.5, D5.6, D5.7


WP6 : Management

Work Package 6 Leader: UniS

This work package will produce annual project and risk assessment, lessons learned, integration, periodic activity reports.

Deliverables : D6.1, D6.2, D6.3, D6.4