Project Marketplace
The project exchange offers you inspiration for exciting projects with BASE topics. Choose one of the project proposals as a basis and develop a real project during the consultation hours!
In addition, ongoing projects can win new members by a) explicitly searching for people with certain skills ("I have project places"), or b) interested students can register on a list of their skills ("I am looking for projects").
Project proposals
InformatiCup 2020
Every year the German Informatics Society organizes a student competition, the InformatiCup (price: 4,000 EUR, in addition students receive ECTS for participation). This year the focus is on the fight against pandemics. The competitors are to implement a web service that plays the "pandemic game". Further information can be found on the official website. Meanwhile it is also advertised accordingly in the trade press.
In other universities the competition is partly deeply rooted. We also want to establish the InformatiCup in computer science and advertise it annually. Interested students can register until the end of the month. The submission deadline is 15 January. Students of computer science/business informatics can obtain 6 ECTS points for participation (registration at base.camp required).
Further information: Flyer for participation
Contact persons: Marcus Soll (GI), Eugen Ruppert (base.camp)
Interactive questionnaire for ethical self-evaluation
Research projects in computer science can raise various ethical questions, e.g. with regard to the collection and processing of personal data. In order to support the independent reflection of the ethical justifiability of certain research projects, the local ethics committee of the Department of Computer Science would like to develop a self-evaluation tool for students and members of the department.
This base.camp project involves the technical realisation of an interactive questionnaire which is available online to researchers interested in self-evaluation. A technical framework and barrier-free design for the online tool will be developed. The design should allow the possibility of adapting content via a content management system (CMS) and ensure long-term maintainability.
We are looking for a motivated student* who can implement the development of such a self-evaluation tool in a timely manner. We also expect openness and willingness to deal with content related questions of research ethics in computer science. If the project starts successfully, an employment as a student assistant may be possible.
Since a continuous support of the system is to be guaranteed, we are looking for applicants* who are not yet close to graduation and who will remain with the department for a little longer.
Target project duration: 6 months / 3-6 ECTS points
Contact persons: Eugen Ruppert (base.camp), Gernot Rieder (EIT)
Speech recognition on embedded systems / Android
Are you curious about speech recognition, deep learning and embedded systems / Android? Have you always wanted to realize an open source project and collect credit points for it at the same time?
Speech recognition is an important component of various speech controlled systems, such as assistance systems. Often, however, data is transferred to the "cloud" or other remote servers. At the same time, the storage capacities and computing power of embedded systems and smartphones are increasing, so that it is now possible to recognize and transcribe speech spoken directly on the device (smartphone, Raspberry Pi, etc.). However, this requires adjustments to model complexity and size on the one hand and a connection to a user interface (e.g. Android App) on the other hand. For such a project we offer:
600+ hours training data for spoken German
Provision of GPUs and workstations via the base.camp
Know-how and assistance with the handling and adaptation of open-source speech recognizers such as Kaldi (http://kaldi-asr.org/) or ESPNet (https://github.com/espnet/espnet)
Sample training scripts for German speech recognition with Kaldi are also already available (https://github.com/uhh-lt/kaldi-tuda-de).
Hardware (e.g. Raspberry Pi) is provided by base.camp.
The goal of the open source project is to create a prototype application that allows to run corresponding models for German in real time on embedded systems, e.g. on Android or with a Raspberry Pi.
Contact persons: Benjamin Milde (LT), Eugen Ruppert (base.camp)
Social media analysis based on the Gab.ai Corpus
On the social media site Gab.ai everyone can freely express their opinion. That is why there is a lot of toxic language, insults, but also discussions about new technologies, politics and society. That is why the corpus offers different possibilities for analysis:
Preparing the development of the platform (posts, users, ...)
Classification of toxic comments
Trend analyses
Network evaluation (by topic, user groups, ...)
The corpus contains more than 30 million posts, which corresponds to about 60GB of data.
Contact person: Eugen Ruppert (base.camp)