MASTER THESIS - NEUROMORPHIC HUMAN MACHINE INTERFACES (M/W/D)
- fortiss GmbH Jobportal
- Bachelor- /Master- /Diplomarbeiten
- Bachelor-/Master-/Diplom-Arbeiten
- IT
Who are we?
fortiss is the research institute of the Free State of Bavaria for software-intensive systems and services with headquarters in Munich. The institute currently employs around 120 employees, who collaborate on research, development and transfer projects with universities and technology companies in Bavaria, Germany and Europe. Research is focused on state of the art methods, techniques and tools of software development, systems & service engineering and their application to reliable, secure cyber-physical systems, such as the Internet of Things (IoT). fortiss has the legal structure of a non-profit limited liability company (GmbH). www.fortiss.org
In the Neuromorphic Computing department, we propose a Master’s Thesis: Neuromorphic Algorithms for Human Machine Interfaces.
The background:
To bring intelligent human machine interfaces to daily life devices, they have to be designed to operate in real-time, on a tight energy budget and in a small form factor. The fortiss Neuromorphic Computing group has been leading efforts to implement human machine interfaces based on neuromorphic sensors, algorithms and hardware. These show huge potential to stay within the latency, energy and size requirements. In recent work, a fully neuromorphic gesture recognition demonstrator was developed. It contains biological-inspired sensing (event-based camera) and brain-inspired processing (spiking neural networks running on neuromorphic hardware). In the follow-up project the system should be expanded to deliver more information about human movements in the form of human poses.
In the Neuromorphic Computing department, we propose a Master’s Thesis: Neuromorphic Algorithms for Human Machine Interfaces.
The background:
To bring intelligent human machine interfaces to daily life devices, they have to be designed to operate in real-time, on a tight energy budget and in a small form factor. The fortiss Neuromorphic Computing group has been leading efforts to implement human machine interfaces based on neuromorphic sensors, algorithms and hardware. These show huge potential to stay within the latency, energy and size requirements. In recent work, a fully neuromorphic gesture recognition demonstrator was developed. It contains biological-inspired sensing (event-based camera) and brain-inspired processing (spiking neural networks running on neuromorphic hardware). In the follow-up project the system should be expanded to deliver more information about human movements in the form of human poses.
Your topic:
So far, an initial spiking neural network (SNN) for the task of human pose estimation from event-based vision data was developed. In this work, the SNN algorithm shall be improved for inference on emerging neuromorphic hardware (Intel Loihi 2). Potential improvements can be made to the SNN architecture, the training procedure and by applying hardware-software co-design techniques (Quantization, Pruning, Knowledge Distillation). The goal of the thesis is the implementation on Intel Loihi 2 and evaluation according to performance benchmarks (accuracy, latency, energy).
Moreover, the developed algorithm can be applied to a sports analysis and/or music generation use-case.
Moreover, the developed algorithm can be applied to a sports analysis and/or music generation use-case.
Your profile:
- Bachelor degree in electrical engineering, computer science, machine vision or software engineering and an excellent track record in your current master’s studies
- Good skills with software development (Python, Git, C++ is an asset)
- A pro at designing object oriented, modular, clean, well-documented code
- Strong experience with PyTorch and/or other ML frameworks
- Strong experience with machine vision (state-of-the-art CNN, deep learning, etc.)
- Knowledge in CNN optimization techniques for efficient inference is an asset
- Preliminary knowledge of Neuromorphic (spiking) architectures is an asset
Our offer:
- Work in partnership with world leading companies in computer technology
- Central Munich, high and modern premises with 360° view on Munich and the Alps
- Access to cutting edge robotics and neuromorphic hardware
- Enjoyable working atmosphere
- Please notice that this Master thesis is not remunerated.
Did we catch your interest?
Please submit your application with a cover letter, a detailed CV, and a current transcript of records.
Job-ID: NC_MS_2023_02
Contact: Michael Neumeier
Job-ID: NC_MS_2023_02
Contact: Michael Neumeier
- Ralf Kohlenhuber
- Human Resources Administrator