Seminar Collective Intelligence: Crowdsourcing
|Dozent(en)||Maria Maleshkova, Fabian Flöck, Maribel Acosta, Kathrin Noack|
|Übungsleiter||Maria Maleshkova, Fabian Flöck|
Unter kollektiver Intelligenz versteht man eine Form der Intelligenz, die durch die Zusammenarbeit und/oder den Wettbewerb vieler Individuen zustande kommt. Die Zahl der Anwendungen, die auf dem Prinzip der kollektiven Intelligenz beruhen, steigt stetig. Bekannte Beispiele für das Ausnutzen kollektiver Intelligenz im Internet sind Wikipedia oder Crowdsourcing-Portale wie Amazon Mechanical Turk.
English is strongly preferred for this seminar, although it is not a strict must. This goes for the presentations as well as thesis papers. Topics 4. and 5. can only be worked on in English.
Topics are covered by teams of 2 students each.
1. Collaboration and Communication Patterns in Open Source communities
Kathrin Noack, IIWR
The focus of the topic is on the communication processes which can be found in Open Source communities which enable a heterogeneous group of individuals to create complex software collaboratively. As a way to approach the topic the communication patterns of a larger Open Source community should be analyzed. In particular the Apache community or the Linux-kernel community are good examples to focus on. The differing communication channels are of relevance as well as the different community structures.
2. Use of Wikis for intraorganizational collaborative knowledge management and collaborative authoring
Fabian Flöck, AIFB
The focus of this topic is the process of how Wikis are used in organizations (most important: businesses) to collect, integrate and manage the knowledge of their employees.
Specifically, it should be assessed what makes this process work and what makes it fail. Specific real life use cases for these best and worst practices should be provided.
Of course, as not all organizations and setups are the same, these practices should be contextualized by three sets of interacting parameters that define the scenario in which the Knowledge management task takes place:
- Technical setup and features
- Usage and content rights
- Social / organizational preconditions
It should be covered as well if KM is the primary purpose of the wikis in question or if this is a side product of another purpose.
Additional aspects to be covered can include:
+ Integration of Wikis with different collaboration tools inside and outside the company
+ Competition and replaceability with other collaborative (authoring) solutions (e.g. Google docs, Etherpad)
3. Software tools for knowledge management in large organizations
Fabian Flöck, AIFB
How do large organizations collect, integrate and manage the knowledge in their organization?
The focus is on software systems that go beyond simple wiki solutions, e.g. SAP, IBM..
What are typical pitfalls, best practices of this type of task in large organizations, especially companies?
What are their advantages and shortcomings in furthering collective intelligence in the enterprise/organization?
What are the currently most popular, most common systems for which kind of organization, task, size, etc.
What makes them useful/popular in their specific context?
What can be covered by big, integrated software solutions and what has to be done with leaner, maybe more open tools (even public ones, like FB, Twitter, Dropbox?)
All questions should be backed up by actual use cases.
4. Social networks and mobile applications: Non-traditional platforms for crowdsourcing
Maribel Acosta, AIFB
Focus: Exploring alternatives to dedicated microtask platforms for specific tasks
Possible solutions: Social networks (Facebook, Twitter), mobile applications (human.io).
- Selection of the crowd: which are required qualifications and composition of the participants?
- Types of tasks suitable for each platform
- Payment/reward methods
- How to avoid spam? is it different from traditional microtask platforms?
5. Choosing the best answer: validation and aggregation of results from the crowd
Maribel Acosta, AIFB
- Identify different types of tasks to be validated and their outputs.
For each task:
- Identify the rule to select the best answer
- Definition of (quality) metrics
- Reliability of the answer
- Time to process the task
- Small experiment: specific use case to try on Facebook, Twitter or discussion group
6. Software system and applications for innovation trough collective intelligence -- harvesting community and social knowledge for triggering new developments and ideas
Maria Maleshkova, AIFB
- the importance of communities and social structures for triggering innovation
- product innovation
- innovation in organization infrastructure and internal organizational workflows
- innovation systems and solutions based on collective intelligence
- application designed for encouraging innovation by facilitating collective intelligence
- supporting collective intelligence-based innovation
7. Forecasting and trend detection based on collective intelligence in online social networks
Maria Maleshkova, AIFB
- trend detection in newsgroups
- trend and pattern detections based on social interaction
- detection of latent patterns and forecasting short- to medium-term trends in twitter posts
- trend detection based on sentiment analysis of posts
- trend detection based on tagging and bookmaking systems
- using social media to predict future trends in markets
- assisting decision making with the help of collective intelligence
James Surowiecki, The Wisdom of the Crowds, Doubleday, 2004, ISBN 0385721706, oder Deutsch: Die Weishet der Vielen, Taschenbuch, Goldmann, ISBN 3442154464
Don Tapscott and Anthony D. Williams, Wikinomics, Portfolie, 2006, ISBN 1591841933, oder Deutsch: bei Hanser, ISBN 3446412190
Generell zu empfehlen, Rest im Seminar.
Das Kickoff-Treffen wird am 15.10.12 um 14:00 am AIFB stattfinden. Es wird einen Zwischentermin zur Präsentation der bis dahin erlangten Ergebnisse am 12.12.2012 geben. Der Endtermin zur Präsentation der Abschlussarbeit (ca. 20 Seiten), die kurz danach abzugeben ist, findet am 04.02.12 statt. Alle drei Termine sind absolute Pflichtermine.
Die Abschlussarbeit macht den Großteil der Note aus (Rest Endpräsentation) und sollte ca. 40-50 Seiten umfassen. Themen werden in 2er-Teams bearbeitet.