This seminar examines new and emerging circuit and system architectures. Some of these architectures aim to improve one or more design parameters in comparison to conventional approaches, e.g. improve computation speed or reduce power con-sumption. Others have been proposed for the implementation of certain new appli-cation classes such as neuromorphic approaches to deep learning. Again other ar-chitectures have been designed specifically for certain manufacturing technologies with their special properties (e.g. inherent unreliability). Current scientific publica-tions from the following fields are intended as topics:
- Stochastic circuits and architectures
- Approximative architectures
- Neuromorphic architectures
- Memristive architectures
- Memory architectures and In-Memory computing
The following requirements are planned for the seminar:
- A 30-minute oral presentation in German or English (student’s choice).
- One set of presentation slides (PowerPoint, Latex, Open Office…).
- An 8-page written summary of the topic in Latex, LNCS-style is obligatory. The summary can be in German or English (student’s choice).
- Attendance at presentations of other participants and active participation in the discussion.
A detailed seminar description will be distributed in the first meeting (preliminary discussion), which contains the exact requirements for the presentation and the written summary. This description will also contain a list of topics, which includes the suggested previous knowledge for each topic (e.g. computer organization, cir-cuit design). The deadlines will also be discussed and decided in this meeting.
If you are interested please contact florian.neugebauer@informatik.uni-stuttgart.de
More information about offered seminars can be found at https://www.f05.uni-stuttgart.de/en/cs/students/seminar-registration/